(Studies in The Early Middle Ages, 15) Wendy Davies, Guy Halsall, Andrew Reynolds - People and Space in The Middle Ages, 300-1300-Brepols (2006)

Ordernr.
061297 P1 KATERN 1
P EOPLE AND SPACE

IN THE M IDDLE A GES, 300–1300
Page 1
Ordernr. 061297 P1
STUDIES IN THE EARLY MIDDLE AGES
Editorial Board under the auspices of the

Centre for Medieval Studies, University of York
Elizabeth M. Tyler (University of York)

Julian D. Richards (University of York)
Ross Balzaretti (University of Nottingham)
V O LU M E 15
Page 2
Ordernr. 061297 P2
P EOPLE AND SPACE

IN THE M IDDLE A GES, 300–1300
edited by
Wendy Davies, Guy Halsall and Andrew Reynolds
with illustrations by
Alex Langlands
H
F
Page 3
Ordernr. 061297 P1
British Library Cataloguing in Publication Data
People and space in the Middle Ages, 300-1300. – (Studies in the early Middle
Ages ; v. 15)
1. Human settlements – Europe – History – To 1500 2. Europe – Social
conditions – To 1492 3. Europe – Historical geography
I. Davies, Wendy, 1942- II. Halsall, Guy III. Reynolds, Andrew (Andrew J.)
IV. Langlands, Alex
307.1'4'094'0902
ISBN-13: 9782503515267
© 2006, Brepols Publishers n.v., Turnhout, Belgium
All rights reserved. No part of this publication may be reproduced,

stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical, photocopying, recording, or otherwise,
without the prior permission of the publisher.
D/2006/0095/68
ISBN: 978-2-503-51526-7
Printed in the E.U. on acid-free paper
Page 4
Ordernr. 061297 P1
C ONTENTS
List of Figures vii
List of Photos x
List of Plates xi
List of Tables xii
Note on Glossary Entries xiii
Acknowledgements xiv
List of Abbreviations xv
Introduction: Community Definition and Community Formation 1

in the Early Middle Ages – Some Questions
WENDY DAVIES
Social Identities on the Macro Scale: A Maximum View of Wansdyke 13

ANDREW REYNOLDS AND ALEX LANGLANDS
Settlement Organization and Farm Abandonment: 45

The Curious Landscape of Reykjahverfi, North-East Iceland
BIRNA LÁRUSDÓTTIR
Geography, Communities and Socio-Political Organization 65

in Medieval Northern Iceland
CHRIS CALLOW
Page 5
Ordernr. 061297 P1
Communities of Dispersed Settlements: Social Organization 87

at the Ground Level in Tenth- to Thirteenth-Century Iceland
ORRI VÉSTEINSSON
Boundaries of Knowledge: Mapping the Land Units 115

of Late Anglo-Saxon and Norman England
STEVEN BASSETT
Mapping Scale Change: Hierarchization and Fission in Castilian 143

Rural Communities during the Tenth and Eleventh Centuries
JULIO ESCALONA
Central Places and the Territorial Organization of Communities: 167

The Occupation of Hilltop Sites in Early Medieval Northern Castile
IÑAKI MARTÍN VISO
The Ending of the Roman City: 187

The Case of Clunia in the Northern Plateau of Spain
ADELA CEPAS
Villas, Territories and Communities in Merovingian Northern Gaul 209

GUY HALSALL
Community, Identity and the Later Anglo-Saxon Town: 233

The Case of Southern England
GRENVILLE ASTILL
Marmoutier: Familia versus Family. The Relations 255

between Monastery and Serfs in Eleventh-Century North-West France
PAUL FOURACRE
Narrating Places: Memory and Space in Medieval Monasteries 275

ANTONIO SENNIS
Populations, Territory and Community Membership: 295

Contrasts and Conclusions
WENDY DAVIES
Glossary 309
COMPILED BY GUY HALSALL
Bibliography 321
Index 347
Page 6
Ordernr. 061297 P1
F IGURES
Cover Illustration: Arnarklettur (‘Eagle’s rock’), a natural marker on a boundary

that divides both properties and communes in West Iceland.
Figure 2.1, p. 15. The course of Wansdyke in relation to shire boundaries in
southern England.
Figure 2.2, p. 23. Sections through East Wansdyke after Colt Hoare, Ancient
History of Wiltshire; Pitt-Rivers, Excavations in Bokerly and Wansdyke Dorset
and Wilts.; Green, ‘Wansdyke: Excavations 1966 to 1970’; and Wessex
Archaeology (unpublished report, 1985).
Figure 2.3, pp. 27–30. The course of the Wansdyke frontier: a) West Wansdyke,
b) the Bath-Mildenhall Roman road, c) East Wansdyke, d) the Bedwyn Dykes.
Figure 2.4, p. 38. Itineraries of Mercian and West Saxon kings up to 871 (after
Hill, Atlas of Anglo-Saxon England).
Figure 3.1, p. 47. Reykjahverfi and surroundings.
Figure 3.2, p. 53. Original farms and possible cottages.
Figure 3.3, p. 55. Cottage ruins — site types.
Figure 3.4, p. 57. Earthworks in Reykjahverfi and Hvammsheiði.
Figure 4.1, p. 66. The Eyjafjörður region, where most of the action of Guðmundar
saga dýra takes place.
Figure 4.2, p. 72. Map to show Hörgárdalur and surrounding valleys.
Page 7
Ordernr. 061297 P1
viii Figures
Figure 5.1, pp. 94–95. Eight study areas representing different settlement types
in Iceland: fjord (Hornstrandir, Dýrafjörður), valley (Öxnadalur, Eyjafjörður),
plain (Úthérað, Ölfus) and coast (Grindavík, Innnes).
Figure 5.2, p. 97. Iceland. Places and regions mentioned in the text.
Figure 5.3, p. 103. Place-name evidence for neighbourhoods in Iceland.
Figure 5.4, p. 110. Laugardælir parish.
Figure 6.1, p. 121. Parishes in the vicinity of Wootton Wawen (Warwicks.) at
their late eighteenth-/early nineteenth-century extent.
Figure 6.2, p. 123. Parishes, chapelries, hamlets, etc. in the vicinity of Wootton
Wawen (Warwicks.), with the probable maximum extent of Wootton
Wawen’s parish also shown.
Figure 6.3, p. 136. The approximate extent of the land granted to the minster at
Wootton Wawen in 716x737.
Figure 6.4, p. 141. The probable maximum extent of Wootton Wawen’s parish
shown in relation to natural drainage.
Figure 7.1, p. 144. North-west Iberia in the tenth-century: the Asturian kingdom
and the Castilian county.
Figure 7.2, p. 150. The district of Ausín in the tenth to eleventh centuries.
(Arrows indicate intrusions of neighbouring districts into Ausín’s. Village-
fission cases are grouped by dashed lines.)
Figure 7.3, p. 156. The Orbaneja valley in the tenth to eleventh centuries.
Figure 8.1, p. 169. Hilltop sites of Northern Castile (sixth–eleventh centuries).
Figure 8.2, p. 174. Territorial organization and hilltop sites in the valley of
Valderredible.
Figure 8.3, p. 175. The territory of Alfania-Paredes Rubias.
Figure 8.4, p. 177. The area of Tedeja.
Figure 8.5, p. 185. Hilltop sites and territories in the district of Pernía.
Figure 9.1, p. 191. Roads, towns and late provincial divisions in Roman Spain.
Figure 9.2, p. 203. Detail of the Clunia area (after TIR, K-30, Madrid).
Page 8
Ordernr. 061297 P1
Figures ix
Figure 10.1, p. 211. The Region of Metz.

Figure 10.2, p. 223. The sixth- and seventh-century phases of the cemetery at
Lavoye, ‘La Haie des Vaches’.
Figure 10.3, p. 225. Distribution of sixth- and seventh-century cemeteries within
the civitas of Metz.
Figure 11.1, p. 241. Places mentioned in the text.
Figure 11.2, p. 245. Single-coin finds from southern England (after Blackburn,
‘“Productive” Sites and the Pattern of Coin Loss in England, 600–1180’,
p.32).
Figure 11.3, p. 247. The north-west part of Oxford in the tenth to twelfth
centuries (after Blair, Anglo-Saxon Oxfordshire, p.156).
Figure 11.4, p. 252. ‘Urban fields’ in southern England (after Darby, Domesday
England, pp.311–13).
Figure 12.1, p. 256. Principal places mentioned in charters relating to
Marmoutier.
Figure 12.2, p. 269. Priories of Marmoutier established in the eleventh century
(after Gantier, ‘Recherches sur les possessions et prieurés de l’Abbaye de
Marmoutier’).
Figure 13.1, p. 277. Main places mentioned in text.
Page 9
Ordernr. 061297 P1
P HOTOS
Photo 1.1, p. 9. The contributors at Lara, northern Spain.

Photo 1.2, p. 11. Castrojeriz, the Pisuerga valley, on the western edge of Castile.
Photo 2.1, p. 19. The Vale of Pewsey looking south from the northern scarp of the
vale, with Rybury hillfort in the foreground. Photo: Andrew Reynolds.
Photo 2.2, p. 22. The contributors walking Wansdyke, Orri Vésteinsson in the
lead. Photo: Alex Langlands.
Photo 5.1, p. 98. Fjord settlement. Dalatangi, E-Iceland. Photo: Oddur Sigurðsson.
Photo 5.2, p. 101. Valley settlement. Fnjóskadalur, N-Iceland. Photo: Oddur
Sigurðsson.
Photo 7.1, p. 146. The Liébana valley, Cantabria.
Photo 7.2, p. 153. Ausín. Photo: Paisajes Españoles, S.A.
Photo 7.3, p. 162. El Castillo, La Muela, and Peña de Lara, from Lara.
Photo 9.1, p. 201. The inscription from Peralejo de los Escuderos (Soria). Photo:
Museo Arqueológico Nacional, Madrid, Spain.
Photo 14.1, p. 297. Arnarklettur (‘Eagle’s rock’), a natural boundary marker
between the farm Hreðavatn in the commune of Norðurárdalur and the farm
Laxfoss in the commune of Stafholtstungur, in Borgarfjörður, West Iceland.
The stone is a part of a lava field that forms the natural boundary between the
communes and properties. Photo: Andrew Reynolds.
Photo 14.2, p. 307. The group at the hillfort of La Cerca, Quintanar de la Sierra.
Page 10
Ordernr. 061297 P1
P LATE
Plate 1, between pp. 48 and 49. Hrísakot ruins in the foreground, marked by a
double earthwork. Above middle is the modern farm of Skörð. The straight
earthwork in the foreground marks the boundary between Laxamýri and
Skörð. Photo: Árni Einarsson.
Page 11
Ordernr. 061297 P1
T ABLES
Table 3.1, p. 59. Settlement names in Reykjahverfi.

Table 5.1, p. 93. Geographical and demographic data for the eight study areas
shown on Figure 5.1.
Page 12
Ordernr. 061297 P1
N OTE ON G LOSSARY E NTRIES
Where a word is highlighted in Bold in the following text, followed by an asterisk,

this indicates that there is an entry in the Glossary for this word.
Page 13
Ordernr. 061297 P1
A CKNOWLEDGEMENTS
All contributors would like to express their gratitude to Barry Cunliffe and Chris
Wickham for continued support of the project, and to the British Academy, the
Institute of Archaeology (Iceland) and Consejo Superior de Investigaciones
Cientificas (Madrid) for financial support. Grateful thanks are due to John Blair,
for permission to use his definition of ‘minster’ in the Glossary and to Alex
Langlands for drawing all the figures. We are grateful to Professor Mark Ormrod,
head of the History Department at the University of York, for funding assistance
with editing.
Page 14
Ordernr. 061297 P1
A BBREVIATIONS
AM The Árni Magnússon Institute, Reykjavík.

Burgos Documentación de la catedral de Burgos (804–1183), ed.
J. M. Garrido Garrido (Burgos, 1983).
Cardeña Colección Documental del Monasterio de San Pedro de
Cardeña, ed. G. Martínez Díez (Burgos, 1998).
CIL Corpus Inscriptionum Latinarum.
DB Domesday Book seu Liber Censualis Willelmi Primi Regis
Angliae, ed. A. Farley, 2 vols (London, 1783).
DB Warwicks Domesday Book, vol. XXIII: Warwickshire, ed. J. Plaister
(Chichester, 1976).
MGH Monumenta Germaniae Historica
SRG Scriptores Rerum Germanicarum
SRM Scriptores Rerum Merovingicarum
SS Scriptores
TIR, K-30, Madrid Unión Académica Internacional (Comité Español), Hoja
K-30: Madrid, Caesaraugusta-Clunia, Tabula Imperii
Romana (Madrid, 1993).
VCH Victoria County Histories.
Wissembourg Traditiones Wizenburgenses: Die Urkunden des Klosters
Weissenburg, 661–864, ed. K. Glöckner and A. Doll
(Darmstadt, 1979).
Page 15
Ordernr. 061297 P1
Page 16
Ordernr. 061297 P1
INTRODUCTION : C OMMUNITY D EFINITION

AND C OMMUNITY F ORMATION
IN THE E ARLY M IDDLE A GES – S OME Q UESTIONS
Wendy Davies
T
he contributors to this book share a sustained interest in the way human
communities used land in the Middle Ages and in the way those com-
munities were in their turn shaped by the landed resources available —
that is, in the literal as well as the symbolic interaction between people and space.
We all have a particular interest in the early or central Middle Ages in Europe, and
some of us in the later Middle Ages too. This period of time has a distinctive place
in landscape and land-use history, falling as it does between the models of pre-
history and the detailed recoverable information on intention that characterizes
the modern age. Research methods are therefore necessarily distinctive. There are
plenty of relevant sources available, but their use is rarely straightforward. We
have some more or less imperfect written descriptions of the exploitation of the
land, very unevenly distributed across time and space. We also have a growing
body of archaeological observation, of settlement, burial and linear features as well
as environmental evidence; although directly relevant, this comes with plenty of
problems of interpretation. Alongside those primary sources, materials of later
date, with judicious use, can be employed to illuminate otherwise uninterpretable
fragments, be they sixteenth-century English ecclesiastical dues, eighteenth-
century Icelandic land registers or nineteenth-century Breton cadastral records.
In the essays that follow, we share a commitment to using all available kinds of
source material and to exploiting the interdisciplinary potential of those materials.
We are all conditioned by the orthodoxies of the last generation and so none
of us would subscribe to pure environmental determinism; however, we certainly
do not share the belief that environment had no impact. A population’s economic
Page 17
Ordernr. 061297 P1
2 Wendy Davies
and cultural relationships to the space it utilized are central to any understanding
of pre-industrial societies, as they remain central to image and identity in our
more economically complex post-industrial world. Hence, it is one of our aims to
clarify the impact of the environment; in order to do so we have worked within
the framework of the strong geographic and climatic contrasts between the
regions of the Spanish meseta and Iceland, and between those two and the
temperate zone of lowland England and northern Francia. The Icelandic material
is especially important, given that the country was settled for the first time in the
ninth and tenth centuries AD.
What follows is an exploration of the relationship between populations, territ-
ory and community membership. There have been many studies of territoriality,
animal as well as human; examples such as the classic paper of Wynne-Edwards on
the territory of the red grouse (most pairs using between 2 and 4 ha of land) come
to mind. Social animals tend to be identified with group territories, witness
primates as well as birds: the home range of a group of gorillas may use 10 km2 of
territory, lemurs less than 1 km2; an average 5.6 pairs of great tits occupy about
10 ha of land in good mixed woodland; and most European readers will be
familiar with the concept of the cuckoo as the stranger — the lone intruder into
group territory.1 Comparison of animal territories, discrete to the species, has
sometimes been made with those of human groups but it is much more usual to
distinguish human territoriality from that of animals, since political and propriet-
ary interests intruded determinants additional to the imperatives of food, shelter
and reproduction. In researching human territoriality there has been a good deal
of work by geographers and anthropologists on the distances people move in the
normal course of their lives. A person’s journeys are certainly significant in respect
of resource gathering; but they are also one of the factors that define community:
the daily range of any person will determine the people with whom he or she will
habitually come into contact and thereby will define one of the communities of
which he/she is a member; his or her annual range, or exceptional range, may
1
V. C. Wynne-Edwards, ‘Self-regulation in Populations of Red Grouse’, in Malthus Past and
Present, ed. J. Dupâquier, A. Fauve-Chamoux and E. Grebenik (London and New York, 1983),
pp.379–91 (pp.384–85); T. H. Clutton-Brock and P. H. Harvey, ‘Primate Ecology and Social
Organization’, in Readings in Sociobiology, ed. T. H. Clutton-Brock and P. H. Harvey (Reading
and San Francisco, 1978), pp.342–83 (p.358); V. C. Wynne-Edwards, Animal Dispersion in
Relation to Social Behaviour (Edinburgh, 1962), p.149; T. Malmberg, Human Territoriality:
Survey of Behavioural Territories in Man with Preliminary Analysis and Discussion of Meaning
(The Hague, 1980), pp.43, 32–33.
Page 18
Ordernr. 061297 P1
INTRODUCTION 3
define others, and so on.2 Much of the theory in fact focusses on the relationship
between productivity of the land and its distance from the worker’s settlement:
at its simplest, the greater the distance the worker had to travel the lower the pro-
ductivity. Comparative studies of mobility between village centres and associated
arable highlight differences in culture even within Europe: in Spain they may be
up to 6 km apart and in Sicily up to 8 km, but in France and Germany 2 km or
less. Stronger differences emerge with comparators beyond Europe, like those
between communities that repeatedly use the same arable fields and communities
that use swidden cultivation, where the population density is much lower and the
individual’s range much wider.3
Much of the above work stemmed from concerns about feeding an increasingly
populated world; much therefore relates to modern, twentieth-century popula-
tions and lacks any sense of a past or of change over time. Nevertheless, it has
informed the arguments of prehistorians in their estimates of the range of hunter-
gatherers, or of the density of settlements and of populations, in a far distant past,
in the quest to reconstruct the economic basis of prehistoric communities and
identify the springboards of cultural change.4 Change over historic time has
primarily been the concern of historical geographers and historical demographers,
working in particular on regular and exceptional journeys, in the context of
urbanization or rural migration or travel to markets; in mid-fourteenth-century
Nottinghamshire the mean distance between markets was, at 5.26 km, lower than
the distance to the arable for some southern Europeans, the contrast emphasizing
2
See R . D. Sack, Human Territoriality: Its Theory and History (Cambridge, 1986), for a useful
overview.
3
The classic work is J. H. von Thünen, Der isolierte Staat in Beziehung auf Landwirtschaft
und Nationalökonomie (Berlin, 1876; first part published Hamburg, 1826), translated in Von
Thünen’s Isolated State, ed. P. Hall (Oxford, 1966). See also P. Haggett, A. D. Cliff and A. Frey,
Locational Models (London, 1977), pp.28–29, 198; T. Carlstein, Time Resources, Society and
Ecology (London, 1982), pp.209–32; M. Chisholm, Rural Settlement and Land Use, 3rd edn
(London, 1979), pp.45–49, 58; C. Geertz, ‘Two Types of Ecosystems’, in Man, Space and
Environment: Concepts in Contemporary Human Geography, ed. P. W. English and R . C. Mayfield
(London, 1972), pp.165–80 (pp.167–74).
4
For example, C. Renfrew, ed., The Explanation of Culture Change: Models in Prehistory
(London, 1973); I. Hodder, The Present Past (London, 1982), especially pp.93–158; D. Green,
C. Haselgrove and M. Spriggs, eds, Social Organization and Settlement: Contributions from
Anthropology, Archaeology and Geography, 2 vols, British Archaeological Reports. International
Series, 47 (Oxford, 1978) — note the assumptions of Part 1, Section 1, ‘Frontiers, boundaries and
group interaction’.
Page 19
Ordernr. 061297 P1
4 Wendy Davies
the speed of economic and cultural change in late medieval England.5 A few
historians have reflected this interest by concentrating on particular aspects of
human territoriality and their potential for stimulating change: riding was usually
the quickest way to move about and the fastest riders could make outstanding
time; the increased use of horse power in the late Middle Ages extended the dis-
tance people could travel and return home at night; so did improvements to the
road system in sixteenth-century England.6 The work of historical demographers
has been particularly important in tracking changing patterns of mobility within
the family: the distance people would travel to find marriage partners in Spanish
Valencia in the early eighteenth century, where 80% of husbands from one area
came from within the same parish*, is comparable to — though more limited than
— that in parts of seventeenth-century English Northamptonshire, where 80%
came from within 5 km, but distances very rapidly extended in the late eighteenth
and nineteenth centuries in England.7
Where there is a pre-twentieth-century dimension to work on human terri-
toriality, it tends to be focussed on prehistory or on the later Middle Ages and
early modern period, often in the latter case on the beginnings of industrialization
and the associated range of social and economic changes. There is relatively little
on the earlier and central Middle Ages even though these issues of changing territ-
orialities have a particular significance at that period. Early medieval western
Europe characteristically had small-scale polities and small-scale social systems;
while political fragmentation was extreme, political change was often rapid. As
larger polities emerged, governments began to use local communities; in so doing,
they sometimes brought them out of a previous obscurity. Determining the
relationship between the local and the governmental is crucial to understanding
the process of politicization: did the larger polity depend on building blocks of
5
T. Unwin, ‘Rural Marketing in Medieval Nottinghamshire’, Journal of Historical Geography,
7 (1981), 231–51 (pp.243–44); the distance rose by 1600 as the character of marketing became
more urban: ibid., pp.247, 251.
6
F. M. Stenton, ‘The Road System of Medieval England’, Economic History Review, 7
(1936–37), 1–21 (pp.16–17); J. Langdon, Horses, Oxen and Technological Innovation: The Use
of Draught Animals in English Farming 1066–1500 (Cambridge, 1986), pp.254, 270; B. P.
Hindle, Roads, Tracks and their Interpretation (London, 1993), pp.65–66.
7
J. Casey, The Kingdom of Valencia in the Seventeenth Century (Cambridge, 1979), p.211;
R . F. Peel, ‘Local Intermarriage and Stability of Rural Population in the English Midlands’,
Geography, 27 (1942), 22–30.
Page 20
Ordernr. 061297 P1
INTRODUCTION 5
pre-existing communities? Or did the needs of government in effect call local

communities into being? Or did both — or other — things happen?
There were of course all kinds of community in the Middle Ages — household,
family, village, district, perhaps even polity, to name the most obvious. The
process of group formation in human society has generated a very large literature,
especially during the nineteenth and twentieth centuries, a literature which has
often treated the emergence of human communities in evolutionary terms. Many
writers have argued that the earliest forms of community arose because people co-
operated as a survival strategy: the theme of social man gradually taming the
surrounding environment is a commonplace in all kinds of theoretical framework;
in such approaches human territoriality is directly relevant to community format-
ion.8 In more recent societies, the close connection with the land has been seen to
weaken, human territoriality to change its nature and the relationships which
underpin society to become commercialized; in classic terms, gemeinschaft (com-
munity) became gesellschaft (society) — small-scale, neighbourhood, ‘genuine’
communities becoming absorbed in large-scale, ‘transient’, competitive society.9
More recent writing reinstates space as a significant factor in social formation, but
a constantly stretching space in an increasingly globalized world; a strong thread
in the most recent writing, as in current politics, is now to do with giving the
spatially defined local group a voice in the widening political community, as it
were ‘real’ community as against ‘imagined’.10 Territoriality has therefore in a sense
returned. In all this there has been a temptation to place medieval society, and
particularly early medieval society, squarely within the models of ‘co-operation
for survival’ and to suppose that most human groups were discrete, and struggling
to make a living, in the medieval world. In this book we start out with the belief
that medieval society, and early medieval society too, was much more complicated
than this, with complex patterns of interdependence and complex relationships
8
R . Leakey and R . Lewin, Origins Reconsidered: In Search of What Makes Us Human
(London, 1992), pp.137–52; for surveys, see M. B. Brewer, ‘On the Social Origins of Human
Nature’, in The Message of Social Psychology, ed. C. McGarty and S. A. Haslam (Oxford, 1997),
pp.54–62 (pp.55–56); P. Dickens, Society and Nature: Towards a Green Social Theory (New York
and London, 1992), especially pp.29–33, 144–46.
9
F. Tönnies, Community and Civil Society, trans. J. Harris and M. Hollis (Cambridge, 2001;
first published 1887 as Gemeinschaft und Gesellschaft).
10
P. Dickens, Society and Nature: Changing our Environment, Changing Ourselves (Cam-
bridge, 2004), pp.21, 225; cf. A. Giddens, Runaway World: How Globalisation is Reshaping our
Lives, 2nd edn (London, 2002).
Page 21
Ordernr. 061297 P1
6 Wendy Davies
with the surrounding natural world. Another central question, therefore, is about
the significance, or even relevance, of simple co-operation for survival in this
medieval world; how far did communities owe their existence to collaboration for
survival and how far did other factors, whether territorial or non-territorial,
condition community identity?
If the communities of the early Middle Ages are clearly underinvestigated,
there has nevertheless been some work which is suggestive. In the last ten years
there has been plenty of attention to community formation in early medieval
Europe in the context of the ‘Transformation of the Roman World’ project.
Much of this work has focussed on ethnicities and on the formation of political
communities.11 Valuable as that work is, it does not help us with the face-to-face
community — the household or the village or the network of associates; it is not
much in touch with the ground. In eastern Brittany in the ninth century, how-
ever, we get a rare glimpse of face-to-face human communities in action — rare
for the period because of the unusual nature of the recording tradition. In this
Breton case the community of the plebs is explicit in the texts that survive —
‘village’ would be the nearest modern analogy for plebs.12 Individuals had mem-
bership of this or that plebs; the members of the plebs, the plebenses, met and did
business together; developments that would affect them were reported to them;
officers were defined in relation to the group. The plebs also had a territorial
dimension, although membership did not extend to everyone who lived on the
territory, which had clearly recognized boundaries: property was located in this
or that plebs. In scale this was usually of the order of 40–50 km 2 , that is 6–7 km
across, twice the size of the later administrative unit of the commune, but it could
be smaller (10–15 km2) or larger (100 km 2), although both extremes are very rare.
Total population size cannot of course be counted at such a date but reasonable
estimates are possible and suggest a relatively high density of population, in the
order of thirty-two people per km 2 (compare fifty-one in 1801 and thirty-eight in
1990). Those residents who were excluded from the community were servile,
often farming a small-holding, but returning all of its surplus to the free peasant
11
See especially W. Pohl and H. Reimitz, eds, Strategies of Distinction: The Construction of
Ethnic Communities, 300–800 (Leiden, 1998); R . Corradini, M. Diesenberger and H. Reimitz,
eds, The Construction of Communities in the Early Middle Ages: Texts, Resources and Artefacts
(Leiden, 2003).
12
For all this, see W. Davies, Small Worlds: The Village Community in Early Medieval Brit-
tany (London, 1988), especially pp.63–85. For some further examples, see now C. J. Wickham,
Framing the Early Middle Ages (Oxford, 2005), cc. 7 and 8; see below, pp.299–300.
Page 22
Ordernr. 061297 P1
INTRODUCTION 7
(or sometimes aristocrat) who was landlord; these people were very clearly pres-
ent; they are named but they had no voice; they did not go to meetings, make
decisions or join in public argument. In the ninth century the plebes were agrarian
communities, with a highly developed sense of the difference between freedom
and servility, living by mixed farming, with a fair amount of arable cultivation.
With such a population density, such an awareness of status, and the cereal-
growing habit, it is not surprising to find communities with a very self-conscious
sense of identity.
Human communities, at least in historic time, are rarely discrete or exclusive
of others. Plebes were not the only kind of rural community in early medieval
Brittany. The community of churchgoers was coterminous with the plebs, but by
implication more populous; the plebs church was central to the social and business
life of the plebs, and several priests served it. Nested within the plebs were family
and other households, and residential units (hamlets) which had identities dis-
tinct from that of the plebs, sometimes called tref or locus13 in the texts (which
sometimes, like Tréal near Ruffiac in the Morbihan, became the focus of a later
commune). It was membership of the plebs, however, which gave ‘civil rights’, to
use an anachronistic analogy, by offering access to mechanisms for the redress of
wrongs and for legitimizing transactions; membership of the plebs had a special
significance: for the purposes of civil society it was the ‘core’ community. Any
individual would, however, have belonged to all or several of these communities,
and to communities beyond.
Although human society was a complex of overlapping communities, plebes
themselves did not overlap: ninth-century texts convey a strong sense of limits;
people moved from one to the next and knew when they were doing so. However,
as well as having communities within them, they were themselves nested within
other kinds of community, some of which were overlapping. Firstly, there were
what one might describe as ‘communities of reputation’. While most of the
plebenses are only ever seen within their own plebs, a small proportion (up to 20%)
went to meetings in two or three plebes, travelling up to 30 km to witness
transactions or serve on panels of jurors.14 These people were trusted peasants,
sometimes wealthy but not necessarily so, whose reputations stretched beyond
their own community’s limits, called in by others to add legitimacy and weight to
proceedings. Then there were the communities of the patronage networks of large
13
Cf. the rather different use of locus in Sennis, below p.276.
14
See Davies, Small Worlds, pp.105–33.
Page 23
Ordernr. 061297 P1
8 Wendy Davies
monasteries like Redon or Landévennec or of the noble retinues of counts and

rulers; these were largely aristocratic or clerical, but the monastic networks could
involve peasants too. Beyond those were the communities of county and diocese
(not usually coterminous) and beyond those again the community of the regnum,
the polity, insofar as it existed. The last three did not have much meaning for
peasant society, but they did for aristocrats and major clerics, who met with count
or bishop or ruler.
The extent to which any of these communities were or are visible on the
ground is debatable, but many years of fieldwork, combined with close attention
to text and enhanced by knowledge of the pre-industrial landscape that is revealed
by the cadastral records of the early nineteenth century, established that the
community of the plebs was in part defined topographically, that is, by large
stretches of uncultivated land (landes) bounding the farmed areas; the landes
themselves were to some extent determined by long ridges of conglomerate and
harder rocks.15 The previously uncultivated landes are often not immediately evi-
dent nowadays, although experience can help identify some of them through soil
type and stoniness; systematic fieldwork can identify others, especially through
the absence of medieval pottery — in this landscape of ubiquitous pottery scatters,
land without sherds of the central and later Middle Ages clearly was not cultivated
at that time.16 These landes (in particular the latter cases) do not always relate to
the underlying geology and the agricultural potential of soil types. They some-
times seem to be determined by cultural factors. The communities of the central
Middle Ages worked so much and so far of the available cultivable land, but not
all of it.17
What determined the size and shape of these ninth-century communities? In
part, but in part only, the underlying geology. They were clearly not defined by
labour patterns: whether one looks at the evidence of 825 or 1825, the arable
occurs in many discrete zones; those who worked it lived beside it; they did not
have to walk to the edge of the plebs or commune because residences, despite the
existence of a nucleated centre, were highly dispersed. The potentially useful tool
of distance between settlement centre and arable simply is not appropriate here
for community definition. The community of neighbourliness, on the other hand,
15
See G. Astill and W. Davies, A Breton Landscape (London, 1997), especially pp.10–32.
16
Astill and Davies, Breton Landscape, pp.120–21.
17
Cf. Carlstein, Time Resources, Society and Ecology, p.293, on the lack of a relationship
between rural population density and environmental factors like soil type or climate.
Page 24
Ordernr. 061297 P1
INTRODUCTION 9
Photo 1.1. The contributors at Lara, northern Spain.
was defined by simple proximity and had more to do with the residential units of
hamlets than with plebes or parishes, a characteristic still evident in the seven-
teenth century when godparents nearly always came from the same hamlet as the
parents of the child baptized.18 All the plebes had centres, and they all had
churches in those centres. Is it possible that the community of the plebs was
defined by the distance of the walk to church? Most churches were within an
hour’s walk of the periphery and even the larger were no more than one and a half
hours’ walk. The answer is not obvious, and argument quickly becomes circular,
but the question is important.
And so, the contributors to this book set out to share experience about
medieval people’s use of space in the areas they know best, to see where there are
common factors and to identify what makes for difference, conscious always of the
varieties of relevant evidence, of the difficulty of marrying different kinds of
evidence, and of the influence that the nature of the evidence has upon interpret-
18
As in Ruffiac in 1670; Davies, Small Worlds, p.128.
Page 25
Ordernr. 061297 P1
10 Wendy Davies
ation. We asked the obvious questions about physical markers in the landscape —
boundary markers, natural and man-made, linear and separate. Were they
present? What did they mark — extents or separations? What of focal points or
central places: were they physical structures like churches or were meeting places
recognizable only through occasions? Were they genuinely focal or were they in
fact incidental to the structures? Did they have practical or symbolic value, or
both, or neither?
We also thought about definitions of community in the Middle Ages and
about evidence of community existence and identity. Who did the defining, the
members of a community or outsiders, contemporary or later? Did the manifestat-
ion of community represent a self-conscious expression of identity or was it
created by some external agency — a district liable to pay tax or tithe, or to
produce ships or fighting men?
We considered issues of extent. Were our communities physically discrete or
scattered? What was their scale; did it relate to the disposition of residences or
labour patterns or walking distance or riding distance; and did this change over
time within our period? Did it include ‘everyone’ in the district, however the
district may have been defined, or only a subset of residents, such as aristocrats or
clerics or fighting men; did it include most but exclude some — such as slaves —
without rights of social participation, faceless residents with no voice? Indeed, did
communities overlap? To what extent were communities ‘nested’ within one
another and how did the elements vary as they scaled up or down?
And we thought about the determinants of community identity. From a wide
range of potential determinants the most obvious are environmental factors, from
physical features and climate to forces for change like farming practice and eros-
ion. Then there are the core demographic factors of population size and density
and the economic imperatives of viable production and necessary exchange. There
are political factors — lordly, governmental, religious, defensive and proprietary.
And there may be cultural determinants — did the circumstance of shared inter-
est, religion, language or approach create community links?
What follows is guided by these questions, addressing sometimes a limited
subset, and sometimes a wider range. Reynolds and Langlands look at a prominent
linear feature in southern England and consider its political context, while Lárus-
dóttir focusses on the interpretation of linear features in Iceland. Callow then
looks at the wide geographical range of communities, and overlapping commun-
ities, in Iceland, followed by Vésteinsson on modern methods of establishing the
existence of communities in Iceland, and the improbability of either settlement
or environment as determinants. Bassett also focusses on method, demonstrating
Page 26
Ordernr. 061297 P1
INTRODUCTION 11
Photo 1.2. Castrojeriz, the Pisuerga valley, on the western edge of Castile.
how the large extent of early parishes in midland England can still be traced;
Escalona too looks at large early units (supralocalities*, with central settlements),
this time in Spanish Castile, adopting some very similar approaches; and Martín
Viso then deals with northern Castile, looking at the changing use of hilltop
sites*, always in some sense central places, whether used by autonomous local
communities or utilized by aristocrats and kings. Still in the same part of Iberia,
Cepas looks at formal, legal definitions of community in Roman Spain, and their
interrelationships; Halsall then concentrates on changing modes of community
definition in Roman to post-Roman Gaul and at the contrast between burial
practice as a spontaneous expression of community and the fiscal district that
simultaneously imposed identity from outside. Astill looks at the very late
acquisition of urban identities in England (in the twelfth century), and explores
the reasons, while Fouracre investigates the vast span of the dependent servile
community of Marmoutier in eleventh-century central France and the overlap
with other kinds of group, before change brought the consolidation of village
communities in the twelfth century. Lastly Sennis draws our attention to
medieval perceptions of space, and their interplay with memory and identity,
Page 27
Ordernr. 061297 P1
12 Wendy Davies
through the eyes of those who made monastic records: in their view monasteries
were in every sense focal, the most central of central places.
We are concerned throughout with how we know what we know, and with
how our perceptions are formed. That provides a common theme running
throughout the book, notwithstanding the varieties of treatment. Although there
may be a temptation for readers to use the Icelandic or the Spanish sections, or
stick to England or France, we hope that they will start at the beginning and read
to the end: there is a logic to the development as it unfolds. If they do this, they
will develop a sense of the comparative, in its many cross-cutting and overlapping
dimensions, both in regional and in thematic terms.
Page 28
Ordernr. 061297 P1
S OCIAL IDENTITIES ON THE M ACRO S CALE:

A M AXIMUM V IEW OF W ANSDYKE
Andrew Reynolds and Alex Langlands
Introduction
T
his contribution considers how medieval populations defined and reacted
to a major political division in the landscape of southern England. We
offer a new look at an old problem: the nature, chronology and context
of one of England’s most substantial yet enigmatic field monuments, its relation-
ship to settlement patterns and social identity. A middle Anglo-Saxon construc-
tion is proposed in the context of territorial disputes between the consolidating
Mercian and West Saxon kingdoms, while the frontier we suggest may well
respect pre-existing territories in places and incorporate earlier features. We argue
also that the nature of this frontier zone markedly affected the way that medieval
settlement patterns emerged from the late Roman landscape.
Individual communities in the Wessex region are historically defined for the
first time in the Domesday Survey of the later eleventh century, a historical source
that identifies two principal social groupings: those at the level of the ‘village’
community, where social stratification is evident from the level of landowner
down to slaves; and the collective community of the hundred* which bound
together individual settlements into a framework of supralocal common interest.
The ‘village’ community can also be approached using physical evidence alone, but
here the definitions of the historian and the archaeologist differ markedly. While
the historian, sources permitting, can address social interaction between indi-
vidual settlements or between households, an archaeological perspective provides
a proximal view based on the study of cultural affinity, economic interaction via
material culture studies, and spatial relationships based on the character, form and
function of settlements. In the absence of explicit documentation relating
Page 29
Ordernr. 061297 P1
14 Andrew Reynolds and Alex Langlands
individual communities to spatial dynamics and territorial definition, collective

participation at varying levels can be approached integrating sources such as place-
name evidence with archaeological material. Our case study approaches people
and space in a region hotly contested at kingdom level, which raises fundamental
issues of how early medieval populations were articulated on the ground.
One of the principal factors of the group’s research was to address ways that
early medieval communities reacted to their environments in greatly differing
landscapes and social and political situations and with varying degrees of internal
and external influences. Community identities are central to the study of frontier
zones, which themselves ensure a dynamic aspect to social definition with scales
of identity running from the individual to the household, and from settlement to
tithing to estate, before social groupings of a yet higher order: hundred, shire* and
kingdom.
Of the regions studied by the group, England was the most intensively settled.
This aspect has left a remarkable record for students of the early Middle Ages both
in physical terms, for there is barely any part of the English countryside that has
not been divided by boundaries at some time, and also in the documentary record
in the form of Anglo-Saxon charter bounds.
This essay offers a ‘revisionist’ view of the Wansdyke earthworks and associ-
ated features, utilizing physical and written evidence. Traditionally the monu-
ment is considered either late or immediately post-Roman (see below). A case is
made here, however, that the frontier’s monumentality, extent and composition
bear closer comparisons with political and military frontiers of middle to late
Anglo-Saxon date. We refer specifically to either West or East Wansdyke, or to
what we term the Wansdyke frontier comprising both stretches of dyke and the
Roman road that links them (Fig. 2.1). A description of the character of the dykes
and related features, including the so-called Bedwyn Dykes, is given below.
West Wansdyke
Wansdyke’s westerly termination is generally considered to be the hill-fort Maes

Knoll in north Somerset, although in the 1920s Major and Burrow noted evi-
dence for its continuation west of the hill-fort.1 Eastwards, between Maes Knoll
and Horescombe in South Stoke parish* (2 km south of Bath), the dyke is
continuous, with the exception of a break of about 3 km where the River Chew
1
A. F. Major and E. J. Burrow, The Mystery of Wansdyke (Cheltenham, 1926).
Page 30
Ordernr. 061297 P1
SOCIAL IDENTITIES ON THE MACRO SCALE 15
Figure 2.1. The course of Wansdyke in relation to shire boundaries

in southern England.
marks the boundary line. Major and Burrow recorded the line of an earthwork to
the north of the River Chew and the first edition OS map of 1811 traces this
feature incorporating earthworks on Publow Hill. At Maes Knoll the ploughed-
out dyke begins as a soil mark 25 m wide. Across open downland at Compton
Dando, English Combe and Stantonbury it maintains a width of 22–25 m, while
a bank 1.3 m high produces a scarp slope of about 8 m.
The dyke faces north on the Oolite ridge between 3 km and 6 km south of the
River Avon flowing from Bath to the Severn Estuary. Through the parish of
Compton Dando the dyke sits on a thin band of Coal Measure, with remaining
stretches on Lias clay formations.2 The dyke crowns open downland at Maes
Knoll at 200 m OD; at Stantonbury and Horescombe it reaches 175 m. Fox and
2
A. Fox and C. Fox, ‘Wansdyke Reconsidered’, Archaeological Journal, 115 (1958), 1–48.
Page 31
Ordernr. 061297 P1
Fox assert that the route chosen is not a commanding one, arguing that the field
of view to the north is obscured by higher ground ‘flanking’ the Avon.3 At no
point north of Maes Knoll, however, are the downs higher than 130 m OD (as at
Newton, St Loe); they normally range between 50 m and 80 m OD.
West Wansdyke shares several attributes of design with East Wansdyke.
Where it can, West Wansdyke crowns the north-facing slope of the highest
ground. This achieves the best defensive and visible line but allows for a series of
strategic high points to be either incorporated into the dyke (e.g. Maes Knoll and
Stantonbury) or sited immediately south of it (hills in Priston parish, 167 m; and
English Combe parish, 143 m). For Offa’s Dyke it is conjectured that beacons
would have signalled from stations at high points behind,4 although little field
evidence has been brought to bear on this issue.
Whether Wansdyke continues west from Maes Knoll to the Severn at Por-
tishead remains contentious. Major and Burrow assembled a series of sketches and
earthwork surveys, while Collinson, vicar of Long Ashton, is said to have recorded
earthworks in Somerset parishes beyond Maes Knoll and as far as Portishead.5
Early in the nineteenth century Colte Hoare met with little success attempting
to trace the same monuments, and little can be added to the debate without new
survey data.6 Fox and Fox dismiss the monument beyond Maes Knoll on the
grounds that west of the river crossing at Saltford (near Keynsham) there is no
need for a fortified boundary.7
The monument recorded by Major and Burrow follows the design specifica-
tions of West and East Wansdyke. It passes across the north-facing slope of
Dundry Hill (202 m) skirting round the highest point of the Oolite ridge (223
m), incorporating earthworks at Windmill Hill and at Portbury Camp on Cony-
gar Hill. This feature is common to the better-evidenced earthworks of West
Wansdyke. Although no references to this dubious stretch of Wansdyke appear
in pre-Conquest boundary clauses, it is noted in two early fourteenth-century
deeds. One, dated 1310, in the reign of Edward II, describes a grant from William
3
Fox and Fox, ‘Wansdyke Reconsidered’, p.45.
4
D. Hill and M. Worthington, Offa’s Dyke, History and Guide (Stroud, 2003), Fig. 46
(p.127).
5
Major and Burrow, The Mystery of Wansdyke; C. S. Taylor, ‘The Date of Wansdyke’,
Transactions of the Bristol and Gloucestershire Archaeological Society, 27 (1908), 131–55.
6
R . Colte Hoare, The Ancient History of Wiltshire, 2 vols (London, 1812–21).
7
Page 32
Ordernr. 061297 P1 KATERN 2
de Goudulph to Adam de Cloptone of a cottage with adjoining land in Aystone

juxta Bristole, on the eastern side of Venelle de Wondesdiche; the other grant is
closely comparable.8 The date of both deeds makes it unlikely that the term was
retrospectively applied, and they surely represent a genuine documentary refer-
ence to Wansdyke along a course previously dismissed.9
The intersection of the Fosse Way with West Wansdyke on Odd Down has
considerable command over lines of approach to Bath, where no fewer than three
Roman roads converge along with the Jurassic way. Bath is a major strategic centre
and, if in the hands of an antagonistic northern aggressor, as it was from a West
Saxon perspective, of fundamental importance to defend against.
The Roman Road from Bath (Aquae Sulis) to Mildenhall (Cunetio)
Opinion is divided on the course of Wansdyke from Horescombe Down to

Morgan’s Hill, the western terminus of the stretch known as East Wansdyke.
Major and Burrow found evidence for a linear earthwork running across the top
of Horescombe Vale on Combe Down. Modern development obscures most of
this possible stretch, although its path conforms to the design specifications of the
better-preserved earthworks on the Bath-facing scarp of Bathampton Down
encompassing Bathampton Camp overlooking Bath at 204 m OD. Major and
Burrow’s findings convinced neither Crawford nor Fox and Fox.10 In 1956–57
Clark set out to identify what relationships existed between West Wansdyke and
the Roman road from Aquae Sulis (Bath) to Cunetio (Mildenhall, near Marl-
borough).11 Rejecting slim evidence for a linear earthwork connecting the two at
Bathford, Clark, like Crawford, opted for a course incorporating the southern
loop of the Avon. Thus West Wansdyke abutted Horescombe Brook in the
Horescombe Vale and adopted the course of Midford Brook into the Avon, where
the boundary would have followed the loop of the Avon up to Lacock where it
would continue its course along the Roman road. Clark assumed that a bank and
ditch would reappear at this junction and continue eastwards to Morgan’s Hill.
His excavations revealed no significant developments to the fabric of the Roman
8
Taylor, ‘Date of Wansdyke’, p.134.
9
Fox and Fox , ‘Wansdyke Reconsidered’, p.37.
10
O. G. S. Crawford, Archaeology in the Field (London, 1953), pp.252–57; Fox and Fox,
‘Wansdyke Reconsidered’, p.36.
11
A. J. Clark, ‘The Nature of Wansdyke’, Antiquity, 32 (1958), 89–97 (pp.92–93).
Page 33
Ordernr. 061297 P1
road, while Fox and Fox investigated the termination of East Wansdyke c.270 m
west of Morgan’s Hill.12
Between West and East Wansdyke the boundary runs over Oxford Clay
through Selwood Forest, a significant boundary to north–south movement. Asser
referred to Selwood as Coit Maur, ‘great wood’.13 That it represented a boundary
of political significance between two halves of the later kingdom of Wessex is
clear. The Anglo-Saxon Chronicle, for example, refers to Aldhelm, as the bishop
‘west of the woods’ s.a. 709, and in AD 893 King Alfred, in countering the Danes,
summoned men ‘both west of Selwood and east’.14
East Wansdyke
Wansdyke is most impressive over the Wiltshire downs and is continuous for 19.5
km from Morgan’s Hill to the western edge of Savernake Forest. On Bishop’s
Cannings Down, including the counterscarp bank, it is over 45 m wide. A bank
of c.5 m produces a scarp slope of 12.5 m. In comparison, Offa’s Dyke reaches a
height of 2.44 m with a scarp slope of 9 m and is 21 m wide.
The ridge of Upper Chalk that bounds the Vale of Pewsey to the north side is
followed by the course of the dyke. The highest points on the downs all lie within
the strip of downland between the dyke and the northern scarp of the Vale of
Pewsey; again comparable to Offa’s Dyke.15
East Wansdyke does not apparently incorporate prehistoric fortifications,
although it runs close to Rybury hill fort near All Cannings. It is possible that Tan
Hill served a role similar to that of Stantonbury on West Wansdyke as virtually
the whole monument from Morgan’s Hill to West Woods is visible from the
summit. In particular, the dyke forms the northern side of a rhomboid-shaped
enclosure of unknown date on Tan Hill.
Fowler’s investigation of access points through East Wansdyke in West Woods
uses charter perambulations. Taking each reference to a geat or gete along the
course of the dyke, Fowler employs earthwork survey to suggest ten original
gateways, although the evidence for ‘outworks’ to the north of several of the gaps
12
13
Asser’s Life of King Alfred, ed. W. H. Stevenson (Oxford, 1904), ch. 55.
14
The Anglo-Saxon Chronicles, ed. M. Swanton (London, 2000).
15
Hill and Worthington, Offa’s Dyke, History and Guide, p.127, Fig. 46.
Page 34
Ordernr. 061297 P1
Photo 2.1. The Vale of Pewsey looking south from the northern scarp of the
vale, with Rybury hillfort in the foreground. Photo: Andrew Reynolds.
in the dyke might just as easily have resulted from breaching the monument.16
This aspect invites comparisons with Hadrian’s Wall and the issue of controlled,
not restricted movement. Regular gates within West Woods alone allow for free
but monitored movement. The pattern of milecastle > turret > turret > milecastle
> turret > turret > milecastle of Hadrian’s Wall is, Fowler argues, echoed in the
spacing of main and minor gates on Wansdyke: main > minor > minor > main >
minor > minor > main.17
Another feature of interest is a possible marking-out line selecting the best
course in relation to local topography and field of view.18 On Offa’s Dyke, during
16
P. J. Fowler, ‘Wansdyke in the Woods: An Unfinished Roman Military Earthwork for a
Non-event’, in Roman Wiltshire and After: Papers in Honour of Ken Annable, ed. P. Ellis (Devizes,
2001), pp.179–98 (pp.187–91); Judie English, pers. comm. (2004).
17
Fowler, ‘Wansdyke in the Woods’.
18
Fowler, ‘Wansdyke in the Woods’, p.191.
Page 35
Ordernr. 061297 P1
a breach of the bank for drainage works, a marking-out ditch was identified.19
Perhaps a year had elapsed between the conception and construction of the dyke
prompting Hill to conclude that it was a ‘mature, planned response rather than
a panic measure’.20
The Ridgeway or Icknield Way runs from the Wash in East Anglia, across
central and southern England into the Vale of Pewsey, and possibly beyond to the
Solent. Fowler’s study identifies a 5 km-broad corridor stretching from Red Shore
to Short Oak Copse where the Ridgeway exists as a series of braided routes
running across varying terrain. Wansdyke was designed, he argues, to marshal
movement along these routes.21
The Bedwyn Dykes
The Bedwyn Dykes run over clay soils better suited to cultivation, and thus
ploughing and erosion. At no point does the width of the dyke exceed 5.5 m.
Stretches of what is better described as a ‘system’ of dykes have been recorded in
the parishes of Great and Little Bedwyn, Shalbourne, Ham and Inkpen.22 Major
and Burrow, working from Colte Hoare’s notes, were content that this repre-
sented a continuation of Wansdyke and identified a continuous course from
Savernake to Inkpen Beacon. Today, the dykes are a series of apparently discon-
tinuous stretches of earthwork of differing proportions.
Sections in Belmore Copse and Chisbury Wood have been discounted by both
Crawford and Fox and Fox.23 While Crawford implicitly refers to the stretch of bank
and ditch that runs in a south-easterly direction from Chisbury Iron Age fort as
‘Wansdyke’, Fox and Fox see no reason to view this as a continuation of Wansdyke.
Although Colte Hoare’s account of the Bedwyn Dykes is early nineteenth cen-
tury, stretches known to the vicar of Ham as late as 1913 had disappeared in his
19
D. Hill, ‘The Construction of Offa’s Dyke’, Antiquaries Journal, 65 (1985), 140–42
(p.141).
20
Hill, ‘Construction of Offa’s Dyke’, pp.141–42.
21
22
E. Hostetter and T. N. Howe, The Romano-British Villa at Castle Copse, Great Bedwyn
(Bloomington, 1997), p.359.
23
Crawford, Archaeology in the Field, p.257; Fox and Fox, ‘Wansdyke Reconsidered’,
pp.18–19.
Page 36
Ordernr. 061297 P1
own lifetime.24 Over one hundred years of agricultural activity could easily destroy
such earthworks on clay soil. Whether the Bedwyn Dykes were continuous will
perhaps never be known, although other evidence suggests its course.
Like Fowler, we can utilize pre-Conquest charters. For this area grants for
Great Bedwyn, Burbage and Little Bedwyn all have boundary clauses (S756, S688
and S264).25 Building on Crawford’s 1942 map of the features in these boundary
clauses, there are nine references to gates, all in close proximity to the dyke as
plotted by Colte Hoare and Major and Burrow. Crawford suggests that these
‘gates’ represent breaks in hedgerows,26 yet it is likely they are related to more
substantial boundaries. Straet Geat, located in the south-east corner of the Great
Bedwyn estate, for example, is the point where the Roman road from Cunetio to
Venta Belgarum (Winchester) passes through a possible stretch of dyke, identified
by Major and Burrow.27
As with the dubious stretch from Maes Knoll to the Severn, reference to the
Bedwyn Dyke in association with Woden occurs in the parish of Ham, where Old
Dyke Lane is referred to on a Common Award map of 1733 as Wans Dyke.28
Dating Wansdyke: A Review of the Evidence
The dating of Wansdyke is less secure than is often presumed. Linear boundary dykes
are widely attested in the British landscape, in both upland and lowland regions,
although the limited attention that they have received relative to other monu-
ments is testimony to the often considerable difficulties of placing them in time
and space. For most writers a late or sub-Roman date is accepted for Wansdyke,
yet this is unproven and arguably reflects a long-running tradition in British field
archaeology of ascribing such a date to substantial linear earthworks not otherwise
placed in the later prehistoric period, such as those on the Tabular Hills of north-
east Yorkshire.29 Elsewhere, linear earthworks like the South Oxfordshire Grim’s
24
Major and Burrow, Mystery of Wansdyke, p.380.
25
Numbers preceded by ‘S’ refer to catalogue entries in P. H. Sawyer, Anglo-Saxon Charters:
An Annotated Hand List and Bibliography (London, 1968).
26
O. G. S. Crawford, ‘The Anglo-Saxon Bounds of Bedwyn and Burbage’, Wiltshire
Archaeological and Natural History Magazine, 131 (1942), 280–301.
27
Major and Burrow, Mystery of Wansdyke.
28
Crawford, Archaeology in the Field, p.257.
29
D. A. Spratt, Linear Earthworks of the Tabular Hills, Northeast Yorkshire (Sheffield, 1989).
Page 37
Ordernr. 061297 P1
Photo 2.2. The contributors walking Wansdyke, Orri Vésteinsson in the lead.
Photo: Alex Langlands.
Ditch, initially claimed as sub-Roman, are also likely to be Iron Age, a date sup-
ported by archaeological excavation.30 Longer chronologies for the construction,
augmentation and alteration of major earthworks should also be given serious
consideration. This factor is exemplified no more clearly than in the remodelling
of Bokerley Dyke on the Wiltshire/Dorset border, a later prehistoric earthwork
subjected to various alterations, including a late phase when the Dorchester to
Old Sarum Roman road was cut through by a massive ditch at the point where it
passes through the earthwork.31
Wansdyke has engaged distinguished antiquarians and archaeologists includ-
ing Pitt-Rivers, Cyril and Aileen Fox, Albany Major and Edward Burrow, O. G. S.
30
M. Hughes, ‘Grimsditch and Cuthwulf’s Expedition to the Chilterns in AD 571’, Antiquity,
5 (1931), 291–314; R . Bradley, ‘The South Oxfordshire Grim’s Ditch and its Significance’,
Oxoniensia, 33 (1968), 1–13; J. Hinchliffe, ‘Excavations at Grim’s Ditch, Mongewell, 1974’,
Oxoniensia, 40 (1975), 122–35.
31
H. C. Bowen, The Archaeology of Bokerley Dyke, ed. B. N. Eagles (London, 1990).
Page 38
Ordernr. 061297 P1
Figure 2.2. Sections through East Wansdyke after Colt Hoare, Ancient History of
Wiltshire; Pitt-Rivers, Excavations in Bokerly and Wansdyke Dorset and Wilts.;
Green, ‘Wansdyke: Excavations 1966 to 1970’; and Wessex Archaeology (unpub-
lished report, 1985).
Crawford and Tony Clark and more recently Peter Fowler; but fundamental ques-
tions relating to the origin of the dyke remain unresolved. John Aubrey first recog-
nized that the western termination of East Wansdyke overlay the Roman Road on
Morgan’s Hill, thus demonstrating its Roman or later origin, at least at that point.
As early as the sixteenth century, Leland commented that the earthwork was
built to separate the kingdoms of Wessex and Mercia, while Aubrey and Stukeley
maintained conflicting views as to its date and origins.32 Colte Hoare traced East
Wansdyke beyond Savernake Forest, identifying stretches of bank and ditch on
Aldermaston common near Silchester, although his search to the east of Inkpen
was fruitless.33
Pitt-Rivers recognized that the only means to establish a chronology was
targeted excavation. His two sections through East Wansdyke were cut at Old
Shepherd’s Shore and Brown’s Barn (Fig. 2.2).34 The first failed to retrieve date-
able finds, although samian ware lay on the ground surface beneath the counter-
scarp. The second section yielded samian ware and a late Roman sandal cleat from
32
The Itinerary of John Leland in or about the Years 1535–43, ed. L. Toulmin Smith, 5 vols
(Carbondale, 1964); J. Aubrey, Monumenta Brittanica ([n.d.]); W. Stukeley, Itinerarium
Curiosum, Centuria, 2 parts (London, 1776).
33
Colte Hoare, Ancient History of Wiltshire.
34
A. H. L. F. Pitt-Rivers, Excavations in Bokerly and Wansdyke Dorset and Wilts. 1888–1891
(printed privately, 1892).
Page 39
Ordernr. 061297 P1
the ground surface sealed beneath the bank, thus providing a terminus post quem.
Taylor identified our terminus ante quem by considering references to the dyke
in tenth-century charters.35 Taylor’s consideration of the relationship between
royal grants of land and the course of the dyke suggested to him a mid-seventh-
century date for its construction.
Two major surveys of the monument in its entirety have been made. Major
and Burrow’s survey does not opt for a single period of construction, but serves
as an overview of all possible periods.36 Sir Cyril and Lady Fox undertook a full
study of Wansdyke in the 1950s having previously undertaken similar surveys of
Devil’s Dyke, Cambridgeshire, and Offa’s Dyke. Their ‘Wansdyke Reconsidered’
identifies ‘minor’ alignments of straight, sinuous and irregular earthwork and
considers their relationship to local topography.37 The article concludes with a
series of bold summary points. The Bedwyn Dykes and East and West Wansdyke,
they argue, are separate earthworks and thus there is no one moment in time when
a frontier containing all three monuments was constructed.38 References in
charter bounds are thought to imply a pagan dedication, and construction dates
for both stretches are related to the pagan Saxon period.39
Prior to the Foxes’ work, O. G. S. Crawford’s commentary on Wansdyke’s
eastern extent was written to counteract the influence of The Mystery of Wansdyke
that he considered a ‘bad book’.40 Crawford began his survey at Maes Knoll and
finished at the foot of Inkpen Beacon on the Wiltshire/Berkshire county boun-
dary. Unlike the Foxes, Crawford considered East and West Wansdyke one and
the same along with the Bedwyn Dyke.41
Contemporary with the Foxes’ survey, A. J. Clark cut two sections through the
Roman road at Spye Park, with a view to ascertaining the relationship between the
dykes and the road that was thought to connect both stretches. Clark found no
evidence for a bank and ditch in conjunction with the road and concurs with the
Foxes’ interpretation.42 Myres summarizes these early contributions and reiterates
35
Taylor, ‘Date of Wansdyke’.
36
Major and Burrow, Mystery of Wansdyke.
37
Fox and Fox, ‘Wansdyke Reconsidered’.
38
39
Fox and Fox, ‘Wansdyke Reconsidered’, p.45. See Hill and Worthington, Offa’s Dyke,
History and Guide, pp.145–47, for critique of the Foxes’ survey of Offa’s Dyke.
40
Crawford, Archaeology in the Field.
41
Crawford, Archaeology in the Field, p.257.
42
Clark, ‘Nature of Wansdyke’, pp.95–96.
Page 40
Ordernr. 061297 P1
the Foxes’ view.43 Myres, however, questions both that the various stretches of up-
standing earthwork are of different periods and the late sixth-century date for the
construction of East Wansdyke. Although he introduces the ‘origins of Wessex’
as a conceptual tool, this is not developed further than the early seventh century.
Subsequently, excavations have been undertaken at Red Shore and New
Buildings, and more recently just south of Marlborough at Wernham Farm in the
modern parish of North Savernake during the laying of a pipeline.44 Various
sections cut through Wansdyke are shown in Figure 2.2. Environmental evidence
suggests that the dyke at New Buildings lay in abundantly wooded surroundings,
whilst on the downs at Red Shore there is evidence of plants characteristic of
pasture.45
The full profile of the ditch at Wernham Farm was not revealed, although oak
charcoal from a layer of flint rubble deposited near the base of the ditch yielded
a radiocarbon date of cal. AD 890–1160 at 2 sigma (BM-2405). It is possible that
the ditch had been cleaned out and further investigation is required. The course
of the pipeline had been diverted to avoid upstanding sections of the bank, thus
eliminating the opportunity to analyse sealed ground surfaces. The date and
section are published here for the first time.
Eagles considers worm action and, citing A. J. Clark in litt., argues that, at a
rate of 4 cm for every ten years, it would have taken some fifty years for the
discarded finds to reach the stratigraphic levels they achieved in Pitt-Rivers’s
excavations.46 Discussion of the Bedwyn Dykes in the publication of the archaeo-
logical research on the Romano-British villa* at Castle Copse, Great Bedwyn,
argues that they are not part of the Wansdyke complex but correspond with late
Roman activity in the region.47 Fowler’s recent survey of East Wansdyke as part
of the Fyfield and Overton Downs project concluded that it was ‘An unfinished
Roman Military Earthwork for a Non-event’.48
43
J. N. L. Myres, ‘Wansdyke and the Origin of Wessex’, in Essays in British History, ed. H. R .
Trevor-Roper (London, 1964), pp.1–27.
44
H. S. Green, ‘Wansdyke: Excavations 1966 to 1970’, Wiltshire Archaeological and Natural
History Magazine, 67 (1971), 129–46.
45
G. W. Dimbleby contributing to Green, ‘Wansdyke: Excavations 1966 to 1970’, p.136.
46
B. N. Eagles, ‘Evidence for Settlement in the Fifth to Seventh Centuries AD ’, in The
Medieval Landscape of Wessex, ed. M. Aston and C. Lewis, Oxbow Monograph, 46 (Oxford,
1994), pp.13–32 (pp.23–24).
47
Hostetter and Howe, Romano-British Villa at Castle Copse, pp.359–67.
48
Fowler, ‘Wansdyke in the Woods’.
Page 41
Ordernr. 061297 P1
Territory, Warfare and Politics: The Context of Wansdyke
There are numerous reasons for placing Wansdyke in a later period based both on
a reassessment of earlier approaches to dating and by examining in detail its
relationship to patterns of settlement and land use from the late Roman period
onwards. Bonney argued that the dyke was laid over existing estates on the basis
that parish boundaries, including several parcels of land described in charter
bounds, bear little if any relationship to the course of the earthwork (Fig. 2.3).49
Indeed, a series of long, narrow estates, typical of downland and vale landscapes
terminate on the high chalk, yet they do not respect East Wansdyke, nor are they
respected by it. Small projections of land belonging to estates both north and
south of the earthwork give every indication of a complex pattern of landscape
development. Bonney argued for a potentially Roman or even earlier date for
these estates with Wansdyke built without regard to their form. In other words,
while there is no logistical sense in a monumental earthwork following a network
of boundaries of local significance, one might expect estates post-dating the dyke
to utilize it as a boundary feature. West Wansdyke exhibits a similar relationship
to parish boundaries.
There are indications that the estates in question are later in origin than
proposed by Bonney. In the first instance, along the course of the Roman road
between the two stretches of dyke, parish boundaries respect the road in every case
except at Chittoe (Wiltshire), a land unit which may have Roman origins and
within which partly lay the small Roman town of Verlucio (Fig. 2.3).50 Clearly the
pattern of estates post-dates the Roman road, yet by how much? Current thinking
places the development of villages and estates (by the twelfth century, parishes)
between the later ninth and twelfth centuries.51 Evidence for this view is derived
largely from charters granting small parcels of land almost frenetically during the
central decades of the tenth century, with far fewer grants of earlier and later
date.52 Archaeology has revealed a similar chronological horizon with regard to
49
D. Bonney, ‘Early Boundaries in Wessex’, in Archaeology and the Lansdcape, ed. P. J. Fowler
(London, 1972), pp.168–86.
50
Bonney, ‘Early Boundaries in Wessex’, pp.176–78; A. J. Reynolds, ‘From Pagus to Parish:
Territory and Settlement in the Avebury Region from the Late Roman Period to the Domesday
Survey’, in The Avebury Landscape: Aspects of the Field Archaeology of the Marlborough Downs, ed.
G. Brown, D. Field and D. McOmish (Oxford, 2005), pp.164–80.
51
D. Hooke, The Landscape of Anglo-Saxon England (Leicester, 1998).
52
D. Hill, An Atlas of Anglo-Saxon England (Oxford, 1981), Fig. 36 (p.26).
Page 42
Page 43
Ordernr. 061297 P1
SOCIAL IDENTITIES ON THE MACRO SCALE

27
Figure 2.3. The course of the Wansdyke frontier: a) West Wansdyke.

Page 44
28
Ordernr. 061297 P1
Figure 2.3. The course of the Wansdyke frontier: b) the Bath-Mildenhall Roman road.
Page 45
Ordernr. 061297 P1
SOCIAL IDENTITIES ON THE MACRO SCALE

29
Figure 2.3. The course of the Wansdyke frontier: c) East Wansdyke.

Page 46
30
Ordernr. 061297 P1
Figure 2.3. The course of the Wansdyke frontier: d) the Bedwyn Dykes.
Ordernr. 061297 P1
the emergence of enclosed ‘thegnly’ residences, with a hall (often aisled), asso-
ciated church and settlement.53
On closer inspection the local picture is more complex than that indicated by
national overview, and for both Wiltshire and Somerset there are eighth- and
ninth-century grants that describe in topographical terms parcels of land that
compare closely with the apparent economic determination behind the estate
morphology of later grants. The likelihood here, then, is that Wansdyke is indeed
later than the estates it traverses, as Bonney suggested, yet the latter are very likely
middle Anglo-Saxon creations at the earliest. A topic that requires further
research is whether late Anglo-Saxon estates were freshly described when granted
by charter or whether they represent instead the component parts of middle
Anglo-Saxon ‘multiple’ or ‘complex’ estates. The latter view, coupled with a later
dating for Wansdyke, raises interesting questions with regard to the origins of
estates and their continuity at a local level either side of the construction of the
dyke in physical and chronological terms.
It is unfortunate that no document earlier than 903 describes either the East
or West sections of the earthwork. Wansdyke is not mentioned explicitly by Bede,
nor is it found in the Anglo-Saxon Chronicle. In 1924 Passmore brought atten-
tion to Gildas’s reference to a turf wall built from sea to sea, while Bede makes a
fleeting but specific reference to an ‘earthwork’, fortified by a series of towers
constructed by Severus, that stretched from sea to sea.54 Fowler takes this to be
consistent with Wansdyke concluding that the frontier was unfinished, while Hill
and Worthington consider its relevance to Offa’s Dyke.55
Entries in the Anglo-Saxon Chronicle relate battles at wodnesbeorg (now
Adam’s Grave) s.a. 592 and 716.56 The battle site itself is the massive Neolithic
chambered tomb just south of East Wansdyke and alongside the Ridgeway.
Identification of the place as that of the battles is substantiated by a reference to
53
A. J. Reynolds, ‘Boundaries and Settlements in Later Sixth to Eleventh-Century England’,
in Boundaries in Early Medieval Britain, ed. D. Griffiths, A. J. Reynolds and S. Semple (Oxford,
2003), pp.98–136.
54
A. D. Passmore, ‘The Age and Origin of the Wansdyke’, Antiquaries Journal, 4 (1924),
26–29; Bede, Ecclesiastical History of the English People, ed. and trans. B. Colgrave and R . A. B.
Mynors (Oxford, 1969), I.5.
55
P. J. Fowler, Landscape Plotted and Pieced: Landscape History and Local Archaeology in
Fyfield and Overton, Wiltshire (London, 2000), p.196; Fowler, ‘Wansdyke in the Woods’,
pp.196–97; Hill and Worthington, Offa’s Dyke, History and Guide, pp.103–05.
56
J. E. B. Gover and others, The Place Names of Wiltshire (Cambridge, 1939), p.318; Anglo-
Saxon Chronicles, ed. Swanton, pp.20–21.
Page 47
Ordernr. 061297 P1
the barrow as wodnesbeorg in tenth-century bounds appended to a (probably

conflated) charter of 825 for Alton Priors (S272).
While the Alton estate is one of those ‘clipped’ by East Wansdyke it remains to
consider the significance of the name and whether the barrow is named from the
earthwork or vice versa. It is plausible on several grounds that the latter is the case.
As noted above, the Anglo-Saxon Chronicle records the barrow as the site of battles
in 592 and 716, events that took place long before its completion, although the
Chronicle is increasingly reliable in its accounts from c.850.57 If we are to accept
the reliability of the earlier event and, indeed, the location of the site, then any
later projected date of construction and naming of the dyke might logically reflect
the major battles fought on the frontier. Equally, the conjoining of a mythical
ancestor with both barrow and dyke might have been made independently.
The first battle was ostensibly fought between Britons and Saxons, while the
later event was a skirmish between Mercians and West Saxons and is more likely
to have actually happened.58 The locality also saw two documented battles in the
early eleventh century when the Vikings were active in the region at Kennet
(1006), on the Ridgeway just south of Avebury, and at Cannings Marsh (1010),
most likely as it was close to Allington (OE Atheling’s tun or ‘prince’s farm’),
perhaps a royal tax collection centre and a ready supply of provisions for the
raiding army.
The name of the barrow is also suggestive of a pre-Christian cult site; the
coincidence of cult and warfare is a matter of not inconsiderable interest. For the
name of an earthwork in a neighbouring county (Somerset) to be reflected by that
of a barrow many miles to the east is a further matter of significance, implying a
strongly ideologically imbued concept of a major frontier. The course of the
Ridgeway north of wodnesbeorg runs along a valley floor known from charter
evidence as wodnesdene and through a break in the earthwork itself termed
woddesgeat (S449; S272). These Woden names, based on those of the barrow and
dyke, underscore the sense of place of this major waypoint for travellers into
southern England, where the Great Ridgeway crosses a boundary of formerly great
importance and where a series of battles had been fought. Effectively, the locale
around wodnesbeorg was an early medieval ‘hot-spot’, its special character deter-
mined by the coincidence of monuments, communications and politics.
57
J. M. Bately, ‘The Compilation of the Anglo-Saxon Chronicle, 60 BC to AD 890:
Vocabulary as Evidence’, Proceedings of the British Academy, 64 (1978), 93–127.
58
Anglo-Saxon Chronicles, ed. Swanton, p.42, n.3.
Page 48
Ordernr. 061297 P1
A seventh-century silver gilt pyramidal stud, inlaid with niello, found ‘near
Adam’s Grave’ is probably a sword fitting.59 It is tantalizing to think that the
barrow might also have acted as a high-status burial place during the seventh cen-
tury as is known in the region at Swallowcliffe Down and Roundway Down, both
of which also reused prehistoric barrows.60 The likelihood that such high-status
burials were also of ‘sentinel’ type, placed at territorial limits, bears serious con-
sideration here and brings to mind the group of seventh-century finds from
Silchester indicative of another high-status burial along the line of our proposed
frontier (see below).61
The place-name Wansdyke or ‘Woden’s Ditch’ has generated much discussion.
In virtually every article published it is either implicit or stated that construction
of the dyke is earlier than AD 600.62 This appears to be due to historical references
to the activities of entrepreneurial Anglo-Saxon princes and the conversion of
Wessex to Christianity from AD 634. Recent studies of early chronicle references
urge extreme caution in using them as a means to construct historical chronolo-
gies.63 It is also apparent that certain pagan place-names might be fabrications of
tenth-century Christian propaganda.64 Aubrey, in his Monumenta Britannica,
remarks how the name caused local people to conclude erroneously that it was
named because the ‘Divell’ erected it upon a Wednesday. Oman’s assertion that,
ignorant of its origins, the Saxons called it after their ancestral god, as a maker of
all things marvellous, is dismissed as naive by Fox and Fox because it is improbable
that such a ‘sacred’ name should be given by the Saxons to an obsolete defence.65
The establishment of the See of Dorchester in AD 634–35 and subsequent
conversion of Wessex has served as a terminus ante quem for the construction of
59
P. Robinson, ‘The Treasure Act’, Trilithon, 46 (2002), 3.
60
S. Semple, ‘Burials and Political Boundaries in the Avebury Region, North Wiltshire’, in
Boundaries in Early Medieval Britain, ed. Griffiths, Reynolds and Semple, pp.72–91.
61
G. C. Boon, ‘The Latest Objects from Silchester, Hants.’, Medieval Archaeology, 3 (1959),
79–88 (p.83).
62
Myres, ‘Wansdyke and the Origin of Wessex’; Bonney, ‘Early Boundaries in Wessex’, p.174;
L. Alcock, Arthur’s Britain: History and Archaeology AD 367–634 (London, 1971), p.349.
63
B. A. E. Yorke, ‘Fact or Fiction? The Written Evidence of the 5th and 6th Centuries AD ’,
in Anglo-Saxon Studies in Archaeology and History, vol. VI (Oxford, 1993), pp.45–50.
64
S. Semple, ‘A Fear of the Past: The Place of the Prehistoric Burial Mound in the Ideology
of Middle and Later Anglo-Saxon England’, World Archaeology, 30 (1998), 109–26.
65
C. Oman, ‘Wansdyke’, Archaeological Journal, 87 (1930), 60–70 (p.69); Fox and Fox,
‘Wansdyke Reconsidered’, p.40.
Page 49
Ordernr. 061297 P1
the dyke. Even Taylor, opting for a mid- to late seventh-century date, suggests that
Christianity could hardly have, by then, become widespread among the West
Saxons, arguing that the name ‘Woden’ would have had perceived influence for
its constructors as the protector of boundaries.66 Major, however, argues that little
evidence exists in northern mythology to associate Woden, or indeed Mercury,
with boundaries.67
The most recent consideration suggests that only in the context of a ‘joke’
would Christian West Saxon builders have named the dyke so. For Fowler, the
attribution of a pagan name ‘Woden’ can only indicate a pre-Christian origin for
the monument yet, as one of us has noted previously, Woden was a ‘live’ figure in
Christian Anglo-Saxon England appearing, as he does, in King Alfred’s genealogy
not as a god but as an ancestor.68 Furthermore, the evidence of the tenth-century
charters shows beyond doubt that clerics producing boundary descriptions were
not troubled with such nomenclature. The name of the earthworks is undoubt-
edly highly significant. They are the only such linear earthworks to be so-named,
an observation that is made all the more striking by their association. That the
name survived into the tenth century indicates at least a memory that the dykes
were once considered as elements of a common boundary line. The reference to
Woden may have been connected with a desire to name the frontier after a heroic
ancestor, deeply rooted in the familial traditions of the West Saxon royal house.
Barbara Yorke, drawing on work by Dorothy Whitelock, has raised the possibility
that the eighth-century King Offa of Mercia may have been christened so as to
associate him with the heroic ancestor of the same name (himself a descendant of
Woden in the so-called ‘Anglian Collection’).69 Yorke also identifies a possible
link between textual references to the heroic Offa in Beowulf and Widsith, with
an inference to the creation of a boundary in the latter perhaps alluding to the
later King Offa’s dyke.70 Parallels with Wansdyke are clear.
The length of the boundary as described by both stretches of earthwork with
the Roman road between also raises major objections to a Roman or sub-Roman
66
Taylor, ‘Date of Wansdyke’, p.152.
67
Major and Burrow, Mystery of Wansdyke, p.133.
68
Fowler, ‘Wansdyke in the Woods’, p.196; J. Pollard and A. J. Reynolds, Avebury: The
Biography of a Landscape (Stroud, 2002), pp.187–89.
69
B. A. E. Yorke, ‘The Origins of Mercia’, in Mercia: An Anglo-Saxon Kingdom in Europe, ed.
M. P. Brown and C. A. Farr (London, 2001), pp.13–22 (p.16); D. M. Whitelock, The Audience
of Beowulf (Oxford, 1951), pp.58–64.
70
Yorke, ‘Origins of Mercia’, p.16.
Page 50
Ordernr. 061297 P1
date. While a much longer frontier can be suggested (see below), even the widely
accepted stretch must relate to an extensive polity. Eagles and Fowler have sug-
gested sub-Roman elites as the builders (see above). The line of the frontier con-
trasts with the suggested Roman administrative geography (civitas* boundaries),
although Eagles suggests that the boundary between the Dobunni and the Belgae
may have run close to the line of East Wansdyke.71 The settlement archaeology of
the corridor of the frontier reveals virtually no evidence for fifth-century occupa-
tion, and not the faintest trace of an infrastructure capable of instituting a
monumental public work traversing southern England from the Bristol Avon to
the Inkpen Beacon and perhaps much further east as originally conceived.
Eagles’s work on fifth- to seventh-century material in the region has estab-
lished a clear division in northern Wiltshire between material culture of Ger-
manic character found to the east of a north–south line running broadly though
Avebury, and finds and place-names of British type to the west.72 Eagles suggests
an extensive territory, based on Bath, itself only 2 km north of West Wansdyke,
while to the east a contemporary unit, potentially that of a group termed the Can-
ningas is proposed for the Avebury region.73 The proximity of West Wansdyke to
Bath and the probability that the Canningas unit included territory on both sides
of East Wansdyke indicates that the monument was built following a post-Roman
phase of medium-sized territories. That East Wansdyke cuts across Eagles’s zone
of cultural contrast rather than following it yet further strengthens a later dating.
Overall, the extent of the Wansdyke frontier suggests that the growth and consoli-
dation of the West Saxon kingdom provides a more plausible historical context
for its construction.
North Somerset, North Wiltshire and Berkshire were hotly contended during
the enlargement of the West Saxon and Mercian kingdoms, and the 716 event is
significant in terms of placing the Wansdyke frontier within this era. There is a
general consensus that Wansdyke is unfinished. Echoing David Hill, Fowler sug-
gests that the earthworks are ‘a mature, planned response’, the need for which may
have subsided during the course of its construction.74 The struggle for territory
between the Mercians and West Saxons provides several ‘windows’ when such a
project might quickly become redundant.
71
M. Millet, The Romanisation of Britain (Cambridge, 1990); B. N. Eagles, ‘Anglo-Saxon
Presence and Culture in Wiltshire’, in Roman Wiltshire and After, ed. Ellis, pp.199–233.
72
Eagles, ‘Anglo-Saxon Presence’, p.208, Fig. 11:2.
73
Reynolds, ‘From Pagus to Parish’.
74
Page 51
Ordernr. 061297 P1
The termination of East Wansdyke at the junction of the boundaries of

Berkshire, Hampshire and Wiltshire at Inkpen Beacon implies a relationship. The
suggested course west of Maes Knoll is used by, or otherwise runs virtually parallel
to, the shire boundary. The continuation of the boundary of Berkshire in an
easterly direction along what would be the ideal line for a continuing bank and
ditch deserves further enquiry. Whilst the ‘Men of Somerset’ and the ‘Men of
Wiltshire’ had been pressed into constructing dykes, perhaps the ‘Men of
Hampshire’ either failed to pull their weight or the need for it had passed. This
interpretation implies that Wansdyke represents military organization based on
the shire system or at least post-dating the shire boundaries (see below).
While we have presented part of our case for a middle Anglo-Saxon date for
the Wansdyke frontier in its entirety, precisely when it came into being and how
long it lasted are questions that remain unanswered. Competition for territory
between the West Saxons and Mercians from the late seventh to the ninth
centuries is complex, yet poorly documented in the grand narrative sources. The
principal type of evidence from which to assess the pace and extent of overlord-
ship exercised in the region by either polity are the series of grants relating to the
foundation or endowment of religious houses, in particular those at Abingdon,
Bath, Cookham and Malmesbury.
The West Saxons controlled Surrey in the 680s (as evidenced by Cædwalla’s
Farnham charter, S235), Kent by 694 (under Ine), along with Sussex between 690
and 710 and apparently Devon by 722, when Ine’s men were fighting the
Cornish.75 Ine granted land at Streatley in Berkshire between 688 and 690 (S252),
although there is no evidence that he ever controlled old West Saxon lands in the
Dorchester-on-Thames region.
A further indication of West Saxon dominance in Berkshire in the later
seventh century is provided by the foundation of Abingdon abbey, apparently by
a West Saxon King Cissa in the late seventh century.76 By contrast, the nunnery
at Bath was founded by a grant made by an underking of the Hwicce, Osric, c.675
(S51). It may be no coincidence that West Wansdyke runs south of Bath and,
arguably, the shire boundary was moved to the north of Bath after it passed into
West Saxon hands in the tenth century. Overall, territorial claims between the
two kingdoms appear less geographically significant along the western part of the
75
F. M. Stenton, Anglo-Saxon England, 3rd edn (Oxford, 1971), pp.68–73.
76
F. M. Stenton, The Early History of the Abbey of Abingdon (Reading, 1913), pp.9–18
and 49.
Page 52
Ordernr. 061297 P1
frontier than further to the east where both northern Wiltshire and Berkshire
witnessed a series of land grants by Mercian kings, including the estate at Long
Newton (Wiltshire), which was given to Malmesbury abbey in 681 (S73). Estates
in Gloucestershire at Somerford Keynes and Tetbury, both close to the border
with Wiltshire, were granted to Malmesbury under Mercian auspices in the 680s
(S71; S1169). Throughout the eighth century the Mercians apparently dominated
Berkshire. Following Offa’s defeat of Cynewulf in 779 at Bensington, the shire
remained under Mercian overlordship for about seventy years.77 We suggest that
this period provides a clearer social and political context for the Wansdyke fron-
tier than arguments presented to date. It remains to be demonstrated whether this
is indeed the case, although the period following Ine’s domination of southern
England and the apparent West Saxon foundation of Abingdon and perhaps,
more precisely, the immediate period following the West Saxon defeat at Bensing-
ton, deserve serious consideration.
By 850 diocesan geography followed the northern boundaries of Somerset,
Wiltshire and Berkshire, while the post-Danelaw reorganizations saw just Wilt-
shire and Berkshire united by the creation of the See of Ramsbury in 909. Viewed
on this scale it is possible to visualize how north Wiltshire and west Berkshire at
least stand out as a cohesive topographical area, defined by downland. Indeed,
David Hill’s map of places visited by Mercian kings up to 871 shows only one
location, Glastonbury (798), south of our projected frontier, while to the north
royal visitations come close to the line (Fig. 2.4).78 Similarly, Hill’s mapping of the
itineraries of West Saxon kings up to 871 shows only Chippenham north of our
suggested frontier (Fig. 2.4).79 Both Mercian and West Saxon kings were present
at Chippenham in 853 for the marriage of King Æthelwulf’s daughter to King
Burhgred of Mercia, following their joint campaign against the Welsh in the same
year (Anglo-Saxon Chronicle s.a. 853). In a diplomatic and administrative sense
our view of the Wansdyke frontier as a major political boundary, however short-
lived, is corroborated, bearing in mind that such distribution maps lack the details
and circumstances of individual visits over a long period.
77
Stenton, Early History of the Abbey of Abingdon, p.50.
78
Hill, Atlas of Anglo-Saxon England, p.83, fig. 145.
79
Hill, Atlas of Anglo-Saxon England, p.83, fig. 146.
Page 53
Ordernr. 061297 P1
Figure 2.4. Itineraries of Mercian and West Saxon kings up to 871

(after Hill, Atlas of Anglo-Saxon England).
Page 54
Ordernr. 061297 P1
The Administrative Background to the Wansdyke Frontier
Charter material, chronicle references and other Anglo-Saxon diplomatic have

been mined regarding the nature of military obligation and drafting labour, and
taxation from the eighth to eleventh centuries.80 Hill’s comparison of existing
earthworks with projected lengths based on the Burghal Hidage assessments
shows a remarkable correlation.81Although of early tenth-century date, the
Burghal Hidage is thought to relate to a scheme of national defence implemented
by King Alfred.82 It relates to the assessment and/or upkeep of thirty-three
fortifications (burhs*) in southern England, broadly demarcating the extent of
West Saxon power between AD 911 and 919.
Attached to each fort is a hidage assessment ranging from 2400 hides* for
Wallingford to as few as 100 for Lyng, Somerset. Of relevance here is the ‘Calcu-
lation’ appended to one surviving version.83 Stating the ratio of men required to
construct and/or maintain stretches of earthwork, the document specifies four
men for the construction of a ‘perch’ of earthwork (English perch = 5.5 yards,
Germanic perch = 5.16 yards). Thus, the projected lengths of fort walls could be
deduced from the hidage assessments granted to each. Hill set about applying this
theory to archaeological reality and met with a surprising degree of success.84 With
the exception of Bath, Chichester and Exeter (most likely owing to the fact that
these had existing Roman walls), the actual lengths of fort walls came within a 5%
margin of error from those projected from Burghal Hidage assessments.85
80
N. P. Brooks, ‘The Development of Military Obligations in Eighth- and Ninth-Century
England’, in England before the Conquest, ed. P. Clemoes and K. Hughes (Cambridge, 1971),
pp.69–84; N. P. Brooks, ‘England in the Ninth-Century: The Crucible of Defeat’, Transactions
of the Royal Historical Society, 5th series, 29 (1979), 1–20; N. P. Brooks, ‘The Administrative
Background to the Burghal Hidage’, in The Defence of Wessex: The Burghal Hidage and Anglo-
Saxon Fortifications, ed. D. Hill and A. R . Rumble (Manchester, 1996), pp.128–50.
81
D. Hill, ‘The Burghal Hidage: The Establishment of a Text’, Medieval Archaeology, 13
(1969), 84–92.
82
M. Biddle, ‘Towns’, in The Archaeology of Anglo-Saxon England, ed. D. M. Wilson
(London, 1976), pp.91–150 (p.124).
83
D. Hill, ‘The Calculation and the Purpose of the Burghal Hidage’, in Defence of Wessex, ed.
Hill and Rumble, pp.92–97.
84
Hill, ‘Burghal Hidage: The Establishment of a Text’.
85
Hill, Atlas of Anglo-Saxon England, p.85.
Page 55
Ordernr. 061297 P1
Burghal Hidage forts are so regularly spaced and accurately reflexive of their
dependent hinterlands that this ‘scheme’ fits with what is known of the adminis-
trative capabilities of King Alfred. Yet, was Alfred drawing from and revising an
existing system? Did this levy, providing the exact ratio of men to fort wall, have
its origins in an earlier period? In Beowulf, for example, Hrothgar summons men
from far and wide to adorn the wallsteads of his palace at Heorot.86 Is this a poetic
echo of the reality of the Burghal Hidage? Beowulf is now promoted as an eighth-
century work again.87
Contemporary with Alfred, Charles the Bald issued a series of capitularies as
part of his Edict of Pîtres (880s) calling for the construction of fortresses and
bridges along with guard duty in his Frankish realms ‘according to the ancient
custom and that of other peoples’. Charles maintained a diplomatic mission in
England headed by the port-reeve of Quentovic and would have been no stranger
to English military activity.88
Charter evidence shows that Mercian kings were exacting a similar levy in the
eighth century.89 The Trinoda Necessitas comprised three common obligations:
service in the fyrd (military service), work on bridges, and the construction of
fortifications.90 Although found in a spurious charter of King Cædwalla of Wes-
sex (AD 680 x 688) the first genuine reference features in a charter of Offa to his
minister Æthelmund (793 x 796, S139). They describe necessary labours from
which no one might be excused.91 These services are first required in response to
the Danish threat in charters of the Mercian Kings Cenwulf (811, S168) and
Ceolwulf (822, S186). The first explicit West Saxon reference to military obliga-
tions is dated 847 in a charter of King Æthelwulf (S298).
Charter evidence suggests that obligations to undertake public works of
defence existed before the reign of King Alfred. To what extent, however, did
Alfred adapt these to suit his system? If the origins of ‘boroughwork’ and the
Trinoda Necessitas lie in the eighth century, as the Mercian charters demonstrate,
can the same be argued for the ratio of four men for every perch? That the Calcu-
86
Beowulf, ed. J. Alexander (Harmondsworth, 1978), p.37.
87
M. Lapidge, ‘The Archetype of Beowulf’, Anglo-Saxon England, 29 (2000), 5–41.
88
D. G. Russo, Town Origins and Development in Early England, c. 400–950 A .D . (Westport,
CN, 1998), p.207.
89
Brooks, ‘Development of Military Obligations’.
90
W. H. Stevenson, ‘Trinoda Necessitas’, English Historical Review, 29 (1914), 689–702.
91
Stenton, Anglo-Saxon England, pp.286–87.
Page 56
Ordernr. 061297 P1
lation is attached to only one of the surviving versions of the Burghal Hidage
suggests that it may be of separate diplomatic identity, unrelated in origin to the
list of forts.92 Hill has demonstrated the merits of isolating the Calculation from
the Burghal Hidage by applying it to Offa’s Dyke.93
Can the Calculation be applied to Wansdyke or to the two sections of earth-
works? What size of labour force and land unit would be required to construct,
for example, the most continuous stretch of earthwork, at 19.5 km, from Mor-
gan’s Hill to New Buildings? This distance equates to approximately 21,500 yards.
With the Calculation in mind, such a distance divides into approximately 3900
English perches (5.5 yards). With four men required for each perch, the requisite
number of hides is 15,600. Burghal Hidage totals for Hampshire (3520), Dorset
(3060), Somerset (2613), Wiltshire (4800) and Devon (1534 with Cornwall,
perhaps an externally enforced but nominal levy), when combined, provide a
grand total for Wessex of 15,527 hides.
Of course, juggling figures can result in a convincing correlation between
defensive monument and territorial context, but what we have attempted to show
is that construction of the earthworks of the Wansdyke frontier required military
organization on a scale equivalent to that of the burghal system of the late ninth
and tenth centuries. When viewed in such a light, Wansdyke seems less isolated
from its kindred artefacts, Offa’s Dyke and the Burghal Hidage forts.
The Development of Settlement in a Frontier Zone
Patterns of settlement provide further valuable insights into the changing nature
of the delineation of communities from the late Roman period. While not a single
Roman-period estate in England can be mapped on the ground, the distribution
of settlements known archaeologically reveals a dense and stratified pattern based
on villas*, some with outlying farms, upland settlements and small towns. With
regard to East Wansdyke, the settlement pattern to the south of the dyke is one
of regularly spaced farms on the chalk scarp facing south but close to the dyke
itself. To the north there is extensive settlement around Avebury, along the
Kennet valley to the east and on the west-facing scarp of the Marlborough
92
A. R . Rumble, ‘Diplomatic Sub Sections’, in Defence of Wessex, ed. Hill and Rumble,
pp.69–74 (p.70).
93
Hill, ‘Construction of Offa’s Dyke’, p.142; Hill and Worthington, Offa’s Dyke, History and
Guide, pp.116–18.
Page 57
Ordernr. 061297 P1
Downs.94 Very few of these sites have revealed early medieval activity and then
only in the form of a few sherds of chaff-tempered pottery which has a fifth- to
tenth-century date range in the region. Although the late Roman farm on Over-
ton Down excavated by Peter Fowler has revealed high-status occupation into the
early fifth century, evidence for the continuity of the Roman settlement pattern
beyond that period is non-existent.95
The lack of influence of Roman towns on the later pattern of urban centres is
illuminating. As described above, at its maximum extent the Wansdyke frontier
arguably runs from Avonmouth near Bristol to the southern boundary of Berk-
shire, which is itself a creation of a top-down planning measure running in a more
or less straight line to the county boundary with Surrey and the River Thames.
One of the most distinctive features of this corridor is the consistent failure of
Roman towns along its length (with the exception of Bath) to develop into post-
Roman towns, as is common in other regions. From Abonae (Sea Mills near
Bristol) through Verlucio (Sandy Lane near Calne), the newly discovered Roman
small town at the foot of Silbury Hill (near Avebury), Cunetio (near Marlbor-
ough), and indeed Calleva Atrebatum (Silchester) on the Berkshire/Hampshire
border, former Roman centres apparently entered terminal decline at the end of
that period. Verlucio, the Silbury settlement and Cunetio failed, yet they were
replaced by three more or less equally spaced small towns by the tenth century.
This aspect demonstrates at least that the physical geography of Roman central
places was fundamentally changed between the two periods.
Major Early Medieval Frontiers
A profitable approach is to consider how Wansdyke compares with the form and
extent of other major frontiers. An obvious parallel can be drawn with the Anto-
nine Wall in southern Scotland, of the mid-second century, a short-lived frontier
built of turf on a stone base.96 Clearly there is a Roman milieu within which such
a boundary feature can be fitted. By contrast, however, the Wansdyke frontier in-
corporates features of a varied character: earthworks, roads and major topograph-
ical features such as watercourses. Early medieval parallels are close to hand.
94
Pollard and Reynolds, Avebury: The Biography of a Landscape, p.154, Fig. 63.
95
Fowler, Landscape Plotted and Pieced, p.104.
96
P. Salway, Roman Britain (Oxford, 1981), pp.194–202.
Page 58
Ordernr. 061297 P1
Offa’s Dyke is no longer seen as a continuous earthwork, but as part of a boun-

dary that incorporated woodland, waterways and other topographical features.97
Similarly, the course of the Danelaw boundary of later ninth-century date is
famously described in the text known as Alfred and Guthrum as running ‘up the
Thames, and then up the Lea, and along the Lea to its source, then in a straight
line to Bedford, and then up the Ouse to Watling Street’.98 Both these latter
frontiers were drawn up on a grand scale as measures designed to define space over
great distances by substantial political entities. Wansdyke fits with the general
character of these frontiers and appears to reflect the mindset of the larger middle
Anglo-Saxon kingdoms with regard to how they separated themselves from their
neighbours. It is worthy of note here that in virtually all other contexts (for
example settlements and burial location) the precise physical definition of space
in social terms became significant again from the late sixth century, but more
commonly during the seventh and eighth centuries.99
What was the social reality of living in a major frontier zone? Did populations
either side of the Wansdyke frontier view themselves as Mercians or West Saxons,
or were they articulated by supralocal communities, such as the hundred? Were
individual settlements considered to be the lowest common denominator?
The nature of early medieval frontiers has been considered from a documen-
tary perspective in a series of European studies pertinent to this discussion.100
Barrow’s essay on the relationship between frontier and settlement in the Anglo-
Scottish border region notes how ‘the Border was an artificial creation, the
product of a series of compromises or bargains negotiated with little regard for
either the inhabitants or even the geography of the region through which its line
was patiently drawn’. The degree to which the line of the frontier fluctuated be-
tween the tenth and mid-twelfth centuries is of interest with major topographical
features determining limits at any given time.101
The field archaeology of the Wansdyke frontier exhibits exactly what might
be expected of such a frontier, particularly with regard to the relationship of the
97
Hill and Worthington, Offa’s Dyke, History and Guide.
98
The Treaty of Alfred and Guthrum, known as AGu I, is printed in The Laws of the Earliest
English Kings, ed. A. J. Robertson (Cambridge, 1922).
99
Reynolds, ‘Boundaries and Settlements’.
100
R . Bartlett and A. Mackay, eds, Medieval Frontier Societies (Oxford, 1989).
101
G. Barrow, ‘Frontier and Settlement: Which Influenced Which? England and Scotland,
1100–1300’, in Medieval Frontier Societies, ed. Bartlett and Mackay, pp.3–21 (p.4).
Page 59
Ordernr. 061297 P1
dykes to estate boundaries apparently ignored when the county boundary between
Berkshire and Hampshire was created. The failure of Roman towns to regenerate
along the Wansdyke frontier finds useful parallels further afield. For example, in
northern Italy in the early Middle Ages, Gian Petro Brogiolo has noted that
territories quickly dominated by their neighbours exhibit a high degree of central
place continuity, while those in regions of longer-term dispute do not.102
One of the most intriguing questions is why the Wansdyke frontier is so
obscure in documented terms. It was clearly intended as a major undertaking by
a polity based to the south, yet with regard to the earthworks, it has left little trace
on the documented administrative geography of the landscapes that it crosses, nor
does it feature explicitly in the cosmology of the more shadowy accounts of sub-
Roman Britain found in Gildas, and then re-related in Bede and the Anglo-Saxon
Chronicle. It can be tentatively proposed that the Wansdyke frontier was the first
of three successive West Saxon frontiers in central and southern England (Fig.
2.1). The second is the Thames, which became relevant, in this context, once
Berkshire and North Wiltshire were safely consolidated into Wessex by the ninth
century. It can be no coincidence that the Thames rather than the Danelaw
frontier appears to be the concern of military planners with regard to the Burghal
Hidage forts along its course. For those that see the Burghal Hidage as simply a
reshuffling and augmentation of earlier arrangements, the defensive nature of the
Thames frontier is perhaps best seen as that succeeding the Wansdyke line once
firmer boundaries became established between Wessex and Mercia. The Danelaw
boundary represents the third major frontier under consideration.
The physicality of the Wansdyke frontier can be read as an articulation of
communities at the level of the embryonic state. For a short period, imbued with
meaning related by name to a heroic ancestor, the boundary line must have
promoted the forging of a West Saxon identity in early England.103
102
G.-P. Brogiolo, ‘Towns, Forts and the Countryside: Archaeological Models for Northern
Italy in the Early Lombard Period’, in Towns and their Territories between Late Antiquity and the
Early Middle Ages, ed. G. P. Brogiolo, N. Christie and N. Gauthier (Leiden, 2000), pp.299–323
(p.310).
103
Thanks are due to the considerable expertise of Mark Bowden and David Field who
walked part of East Wansdyke with us and provided valuable insights and observations. Andrew
Lawson kindly allowed us to publish the results of the 1985 cutting through East Wansdyke at
Wernham Farm by the then Trust for Wessex Archaeology, while Amanda Grieve of the Ancient
Monuments Laboratory provided information about the radiocarbon date. Judie English and
Graham Brown are thanked for their discussion of the Wansdyke problem.
Page 60
Ordernr. 061297 P1
S ETTLEMENT O RGANIZATION AND

F ARM A BANDONMENT: T HE C URIOUS L ANDSCAPE
OF R EYKJAHVERFI, N ORTH -E AST ICELAND
Birna Lárusdóttir
Introduction
I
n Iceland, relatively little research has taken place in the field of settlement
patterns and the structure of early communities. Yet, for many reasons,
Iceland is an exceptionally good place to study them. The country was not
settled until late in the ninth century and it is generally accepted that the locations
of individual farms have not changed dramatically since then. The Book of Settle-
ments, a twelfth-century source that contains information on the first settlers and
their choice of land, mentions some six hundred farms of which the vast majority
are still inhabited today. Other written sources, such as sagas, are helpful when it
comes to observing how settlement took place. Although these written sources are
limited, one must object strongly to the assertion that Iceland has no prehistory.
The Icelandic landscape has not been as dramatically affected by modern
human impact as many other European countries. Still it is not until recently that
archaeologists have begun to observe the archaeological landscape through field
survey. Adding that to the historical information, a new perspective can be
obtained on the size and shape of communities, how they were formed, and how
they may have changed over time. Changes in settlement patterns in Iceland have
often been explained through environmental determinism, invoking such things
as climatic change, volcanic eruption and erosion. In the following piece a case
study of the formation of a rural community will be introduced and an attempt
made to explain the establishment and disintegration of a small community, using
different types of sources.
Page 61
Ordernr. 061297 P1
46 Birna Lárusdóttir
Field survey in Reykjahverfi in north-east Iceland has recently revealed an

unusual number of ancient-looking farm sites, many of which were abandoned
long before the beginning of the eighteenth century and some, judging from the
shape and layout of the ruins, even as early as the twelfth century.1 This evidence
provides a new perspective on settlement in the area and how it progressed. This
essay first presents an overview of how the area may have been settled, based on
written sources such as The Book of Settlements, sagas of Icelanders, property value
figures, and the location of churches and chapels*. The methods used are mostly
based on the work and theories of Orri Vésteinsson.2 Second comes a presentation
of archaeological sites, some of which can be connected to written sources, while
others were previously unknown. This is followed by a study and comparison of
settlement names. Third, the location of shielings* and some other results from
the survey are explored to cast light on the use of outfields and how outfield
management can relate to the settlement patterns. Finally, an extensive system of
earthworks stretching over a large area in the north-east is discussed to help
explain how settlement in the area progressed.
Reykjahverfi is a shallow valley, the easternmost of three valleys stretching south
from Skjálfandi bay (Fig. 3.1). West of Reykjahverfi lies Aðaldalur, a wide, grassy and
well-defined valley with high-quality land. The Laxá, a river rich in salmon, flows
through it. Although Reykjahverfi is a valley in the sense that it has a river running
through its middle, it is, when compared to the neighbouring valleys, more like a
plain, especially at its mouth, close to the sea, where the land is flat, sloping very
gradually from the bottom up to the edge of the heath. The two valleys, Aðaldalur
and Reykjahverfi, are separated by Hvammsheiði, a heath that stretches from
north to south. Half of Hvammsheiði belonged to Reykjahverfi, at least in later
times, and was defined as common land. It is separated from the farms’ outfields
by a relatively small river. The farms in Reykjahverfi that are inhabited today are
more or less evenly spread in a row alongside the foot of Reykjaheiði heath.
It is important to note that the archaeological remains in this area are better
preserved than in most others, both because of relatively little soil accumulation
and also because human impact has not been as extensive as in many other parts
of the country — the making of hayfields has been more or less centralized around
those farms which are still inhabited, leaving large tracts untouched by bulldozers.
1
B. Lárusdóttir and E. Hreiðarsdóttir, Fornleifaskráning í Reykjahreppi, 2 vols (Fornleifas-
tofnun Íslands, 2002–03).
2
O. Vésteinsson, ‘Patterns of Settlement in Iceland: A Study in Prehistory’, Saga-Book of the
Viking Society, 25 (1998), 1–29.
Page 62
Ordernr. 061297 P1
SETTLEMENT ORGANIZATION AND FARM ABANDONMENT 47
Figure 3.1. Reykjahverfi and surroundings.
Page 63
Ordernr. 061297 P1
Before delving deeper into the archaeological landscape in the area it is

essential to look at the history of its settlement. This survey is based on several
sources: the valuation of farms, location of churches and chapels, boundaries
between farms, and written sources such as The Book of Settlements, sagas and
charters.
History of Settlement
Traditionally, the region is well known in the context of the early habitation of
Iceland. According to The Book of Settlements a man called Garðar Svavarsson
arrived in Húsavík before the actual settlement started, built a house and
overwintered there.3 Although The Book of Settlements is not to be taken literally
it is now considered an important indication of how colonization could have
taken place. It is easy to imagine that Húsavík played an important role in the
occupation of this part of the country as it has the best landing place in the whole
bay of Skjálfandi. As Orri Vésteinsson has pointed out, Húsavík is likely to have
been settled early and those who lived there had every opportunity to influence
the settlers that came afterwards.4 It is likely that the eastern area of Skjálfandi was
inhabited before Reykjahverfi, and the same applies for Aðaldalur, simply because
of quality of land. The greatest part of Reykjahverfi lies higher up, has less arable
land and is therefore not as ideal for an initial colonization.
The only settlement in Reykjahverfi mentioned in The Book of Settlements is
Skörð, founded by a man called Höskuldur. His claim stretches over most of the
study area.5 Skörð had the highest tax value in the area, forty hundreds, along with
two other farms, Laxamýri and Reykir. These farms are all very likely to have been
settled early, simply because they had most of the best quality land. It can even be
argued that Skörð was the least desirable farm of the three, having access neither
to salmon fishing like Laxamýri, nor geothermal heat like Reykir. The River Þverá
flows almost exactly midway between Skörð and Reykir and was probably the
original boundary between the two settlements. There is no obvious natural
boundary between Skörð and Laxamýri but a quite massive earthwork was built
there, presumably during the first ages of settlement. Laxamýri was at first simply
3
Landnámabók, ed. J. Benediktsson, Íslenzk fornrit, 1 (Reykjavík, 1968), p.35.
4
E. Hreiðarsdóttir and O. Vésteinsson, Fornleifaskráning á Húsavík [FS166–01171] (Forn-
leifastofnun Íslands, 2002), pp.7–9.
5
Landnámabók, ed. Benediktsson, p.276.
Page 64
called Mýri (meaning ‘wetland’) but this later became Laxamýri, incorporating the
river’s name, to distinguish it from other Mýri farms. The name Mýri derives from
the rich meadow west of the farm, close to the river Laxá, a strong reason to sus-
pect early settlement there. Late in the nineteenth century, a hard-working farmer
had an impressive irrigation system built, damming the river so its rich nutrients
fertilized the meadow, increasing its productivity. Laxamýri has other outstanding
natural resources, such as salmon fishing in the Laxá river, eggs in nesting colonies
in the islets of Lake Mýrarvatn and seal hunting in the river’s estuary. The farm
Reykir is located just north of a geothermal area with a few hot springs, giving the
settlement its name, Reykir meaning ‘smoke’ — the word settlers used to describe
steam. A small, warm river with extremely grassy banks runs from the springs next
to the farm’s homefield, ideal for letting flow over fields and meadows. A large
part of an irrigation system that was used in the nineteenth and twentieth cen-
turies is still visible south of the farm. As Orri Vésteinsson has shown in his
discussion of Icelandic settlement patterns, natural meadows like these were an
important factor in determining the location of initial dwellings.6
The three farms are all mentioned in early written sources. First and foremost
is the Ljósvetninga saga, which makes Skörð the home of a great chieftain, Ófeigur
Járngerðarson.7 He is also introduced in the Reykdæla saga, and furthermore a
man called Þorkell is said to live in Ytri Skörð or Outer-Skörð.8 Mýri and Reykir
are mentioned briefly in the tale of Vöðu-Brandur.9 All this points to a long-
established settlement in these farms, but at least three or four other farms in the
area seem to have been settled early as well: Saltvík, Skógar, Þverá and probably
also Heiðarbót, although the name ‘-bót’ could indicate a somewhat later date,
deriving from the verb bæta which means ‘to add’. These are the farms with the
second highest evaluations in the area, all of twenty hundrað* (hdr.) except
Heiðarbót, which was only ten hdr. However, these farms are not mentioned until
later: Saltvík in 1296, Skógar in 1449, and Þverá in 1544.10
6
Vésteinsson, ‘Patterns of Settlement’, pp.7–8.
7
Ljósvetninga saga, ed. B. Sigfússon, Íslenzk fornrit, 10 (Reykjavík, 1940), p.3.
8
Ljósvetninga saga, ed. Sigfússon, pp.212–13.
9
Ljósvetninga saga, ed. Sigfússon, pp.125 and 130.
10
Diplomatarium Islandicum eða Íslenzkt fornbréfasafn 834–1600, 16 vols (Copenhagen and
Reykjavík, 1853—1976), II, 316; V , 40; XI, 344–45.
Page 65
Ordernr. 061297 P1
Another indicator of antiquity is the location of chapels.11 In Reykjahverfi

there were chapels in Saltvík, Laxamýri, Skörð and Reykir. A mound called
Kirkjuhóll (‘church mound’) lies close to the farm Heiðarbót. According to the
farmer, over recent years the farm dogs have been caught carrying human bones
from here, quite a clear indication of a chapel site. The only chapel that ever came
close to being a parish* church was in Reykir. In charters of the church at
Grenjaðarstaðir from 1318 and 1394, the possibility that Reykir church might
have its own priest is mentioned, but if not the parish was to be served by an extra
priest at Grenjaðarstaðir.12 If Reykir ever was an independent parish it must have
been subsumed by other larger parishes at an early date.13 In later times the farms
in the area belonged to two parishes: Grenjaðarstaðir and Húsavík. According to
a farm survey of 1712 most of the inhabitants in Reykjahverfi had to travel a great
distance to go to church, which was seen as very unfortunate.14
Yet more evidence for ancient settlement are pagan graves, which until
recently were unknown in the area. In the summer of 2003 a small excavation was
carried out close to an ancient-looking cluster of ruins in Saltvík. It revealed the
remains of at least two pagan graves that had been robbed before 1477, as they
were clearly sealed by volcanic tephra* from that year. Two other discoveries of
human remains should be mentioned. A human skeleton was found a short
distance north of Reykir in the 1960s, and in the 1970s the remains of a skeleton
were unearthed close to the old farm mound in Þverá — oriented north–south.
This could support the idea that Þverá was settled early, that is, before the
conversion in the year 1000.
The name of the area, Reykjahverfi, has a long history. It is first mentioned in
the two sagas that take place in north-east Iceland: the Ljósvetninga and Reykdæla
sagas. In the first, the tale of Ófeigur, a man named Þorbjörn is said to live in
Reykir in Reykjahverfi, which is then obviously the name of a larger area around
the Reykir farm.15 The word hverfi* occurs quite frequently in Icelandic place-
names although it has not been studied specifically. Essentially, it means ‘cluster’
11
Vésteinsson, ‘Patterns of Settlement’. Vésteinsson has also written about parish structures
in Iceland: O. Vésteinsson, ‘Íslenska sóknaskipulagið og samband heimila á miðöldum’, in Íslenska
söguþingið 28.–31. maí 1997: Ráðstefnurit, 2 vols (Reykjavík, 1998), I, 147–66.
12
Diplomatarium Islandicum, II, 433; III, 581.
13
Elsewhere in this volume Orri Vésteinsson discusses how parishes represent communities
in Iceland.
14
Jarðabók Árna Magnússonar og Páls Vídalíns, 13 vols (Copenhagen, 1913–90), XI, 245–50.
15
Ljósvetninga saga, ed. Sigfússon, p.117.
Page 66
Ordernr. 061297 P1
of farms, although it can also mean a valley or a natural depression of some kind
— which can apply to this area. A brief survey of -hverfi place-names shows,
however, that a vast majority of them have nothing to do with natural features but
rather indicate the nature of the settlement. The term, however, is not always
applied in the same way. A common usage is found in fishing areas where a cluster
of households has been formed around a fishing station — for example, Járnger-
ðarstaðahverfi in the south-west. The nearest example of a similar use of -hverfi
in the north-east is Skriðuhverfi in Aðaldalur, where a cluster of farms has been
built around Skriða, a farm that was either dominant from the beginning or later
became dominant. Another example is Kelduhverfi, north-east of Reykjahverfi,
which has become the name of a whole district with many farms; although this is
unusual. In such cases the prefix most often includes a natural element, either a
feature that dominates the area or which describes it. An example of this is
Fljótshverfi (fljót meaning ‘river’) in the south-east, flanked by rivers that were
great barriers in the past. Although this could be the case in Reykjahverfi (Reyk-
or Reykja- being a very common prefix in place-names that derive from geo-
thermal areas, even if no farm with the name is present), it seems more likely that
Reykjahverfi takes its name from the farm Reykir and the surrounding farms and
later expanded to the whole area. It is not very clear from the descriptions in the
two sagas how far Reykjahverfi extends. In the Reykdæla saga a group of men on
their way up north are said to have stayed overnight in Reykjahverfi with a man
called Geirrekur, which seems inaccurate had the area extended as far as it does
now.16 Reykjahverfi seems rather to be the name of a specific location, not a valley
or a whole area. In the farm register of 1712 only four farms are said to be within
the hverfi: Skógar, Heiðarbót, Þverá and Reykir.17
Nowadays and at least since late in the nineteenth century, Reykjahverfi is said
to reach from Skörð south to Geitafell, which is a farm 7.3 km south of the modern
boundary of the district now called Reykjahreppur (Fig. 3.1). At least as early as
the beginning of the eighteenth century, this southern part of the hverfi — more
than one-third of its whole length — belonged to the parish of Grenjaðarstaðir
and now belongs to the district of Aðaldælahreppur. In the beginning of the
eighteenth century this area belonged to the church in Grenjaðarstaðir which used
it, along with other farms, mainly for shieling purposes; so evidently there was
some vested interest at stake. Geographically the boundary line drawn between
16
17
Jarðabók Árna Magnússonar og Páls Vídalíns, XI, 249.
Page 67
Ordernr. 061297 P1
the two districts seems highly unnatural, as it cuts across the narrow end of a valley
(Fig. 3.1). This indicates that the southernmost part of the hverfi was lost to
Grenjaðarstaðir at some point. Thus it seems that Reykjahverfi was originally only
the name of Reykir and perhaps its surrounding cottages, but later stretched out
to a much larger area.
Cottages/Ruins
Reykjahverfi exhibits an exceptional number of assumed farm ruins that consist,

in most cases, of several structures and a homefield boundary (Fig. 3.2). Most of
them are not mentioned in sources until the 1712 farm survey, according to which
the vast majority had been abandoned for a long time. Four were occupied until
the seventeenth century: Brekknakot and Holtakot, belonging to Reykir; Skarða-
gerði close to Skörð; and Maríugerði in Laxamýri which is unknown today or has
become confused with other sites. Three deserted cottages around Reykir are
mentioned in a 1525 register of farms belonging to the bishopric of Hólar, but
apart from that there is little, if any, knowledge of when or why the cottages were
abandoned.18 It is not uncommon for the farm register to mention a few deserted
cottages in each district, but the number in Reykjahverfi is unusually large. It is
important to bear in mind that certain locations referred to as abandoned
farmsteads in the register do not look convincing as such when surveyed in the
field. Sometimes locations seem unfavourable and in other cases ruins are simply
too small and insignificant.
When field survey was undertaken in the area in 2002 and 2003, it soon
became clear that Reykjahverfi has an unusual number of ancient-looking ruins
that appear to be farmsteads. Some of these are mentioned in the 1712 farm
survey, but some were pointed out by local farmers and others seem to have been
completely unknown. A number of sites have been damaged, for example by road
construction or the levelling of homefields, and the locations of a few mentioned
in earlier sources are today unknown. The first question to be answered is whether
all these sites were in fact farmsteads. This is not a simple matter, and probably
requires extensive excavation. However, the majority seem to fit well with the
conventional definition of a farmstead — a unit that consists of several structures:
a dwelling, some outhouses, and preferably a homefield boundary. With one or
two exceptions all seem to fall within that description.
18
Diplomatarium Islandicum, IX , 301.
Page 68
Ordernr. 061297 P1
Figure 3.2. Original farms and possible cottages.
Page 69
Ordernr. 061297 P1
There are two main types of sites (Fig. 3.3): those enclosed by a simple boun-
dary, and more complex ones fenced off by a double or even triple boundary. The
simple ones are generally smaller although they do not necessarily have fewer
structures. At this point, it is not possible to say whether this is an indication of
lower status, possibly different function, or is perhaps related to different speciali-
zation in animal husbandry, chronology or length of occupation. In two cases the
simple boundaries do not enclose the area around the sites, possibly because of
erosion in the case of Anonymous 4. The second example is the site of Litlu-
Skörð, located within a very large boundary that seems to define a territory
around the farm Skörð. Although it has no boundary the Uxastaðir site is also
included here as it is fenced off naturally by lava fields.
The multiple field boundaries are curious and at least four of the sites in that
category share a very similar morphology: one boundary fences off the area around
the dwelling and some outhouses, and an outer boundary defines a space around
the first boundary and seems to have some activity linked to it. This activity is
most commonly revealed by small enclosures of some kind but sometimes by
structures, possibly outhouses. Where there is a third boundary it is not always
very visible, but remains of such features can be seen in at least four cases and
usually mark out a space with no structures within it. Two sites — the ruins in
Íradalur and Skarðagerði — have an area fenced off next to the homefield without
any visible structures. This is somewhat easier to explain than the multiple enclo-
sures; it seems practical to have such an area where ewes or horses, for example,
can be corralled and the enclosed land thus fertilized to extend the homefield.
Similar enclosures were common in the late nineteenth century, called nátthagi
(‘night enclosures’).
Multiple boundaries in Iceland have not been studied before although they
have aroused curiosity in the past. At least two interesting interpretations exist in
Icelandic literature. One comes from an eighteenth-century agricultural treatise
composed as a dialogue or tutorial between a young man and an older farmer. This
source suggests that in the beginning, when the farms were settled, people fenced
off relatively small homefields but the size increased as the number of livestock
grew.19 According to this the multiple boundaries are a sign of how realistic the
forefathers were and how they gradually enlarged their homefields as they were
able to sustain larger numbers of animals. The other explanation, from the early
19
B. Halldórsson, ‘Atli’, in Rit Björns Halldórssonar í Sauðlauksdal (Reykjavík, 1983),
pp.49–213 (p.118).
Page 70
Ordernr. 061297 P1
Figure 3.3. Cottage ruins — site types.
Page 71
Ordernr. 061297 P1
twentieth century, maintains the reverse: that the first settlers fenced off large
homefields around their farms but as time went by and land degraded they were
forced to reduce in size and thus build a new boundary within the limits of the old
one.20 The latter explanation is partly based on the fact that the numbers of cattle
seem to have decreased soon after the end of the Commonwealth period, and
therefore it is a reasonable assumption that less hay from the homefield was
required.21 This interpretation also echoes the romantic vision, inspired by sagas,
which scholars and the public shared in the nineteenth and twentieth centuries.
In this view, the Commonwealth period was seen as the golden age of Icelandic
history, which went into decline after the end of the thirteenth century.
The possibility that such multiple boundaries are contemporary has not been
considered, probably because it does not seem to make much sense. If a farmer
wanted to enlarge his homefield it would seem more sensible to choose the best
land outside it, rather than enclosing an equidistant ring all around it. In addition,
building a new boundary all around the old homefield demands more work than
simply enlarging it in one direction.
So far two sets of double homefield boundaries have been trenched in the
district of Mývatn, south of Reykjahverfi, and in neither case can they be proven
to be contemporary.22 There is as yet no feasible explanation for the function of
such features. In many cases, the innermost boundary stands higher than the
others, simply because it has been rebuilt in connection with later phases of
activity on the site, most commonly linked to the housing and winter grazing of
sheep. However, there are some examples where it is possible to see the order in
which the boundaries were built. In Heiðarbót there is a system of surviving
earthworks just outside the modern homefield, fencing off an area of about 370
x 300 m (Fig. 3.4). The system is very similar to the double- or triple-complex
homefield boundaries mentioned previously. It seems clear from the context that
the innermost boundary came first; secondly a long linear boundary, possibly a
pasture boundary, was added to it, stretching to the north; and thirdly a double
boundary and complex enclosures were added. At this point there is of course no
way of saying whether this happened over a period of ten or two hundred years.
20
Þ. Thoroddsen, Lýsing Íslands, 4 vols (Reykjavík, 1908–22), I, 91–92.
21
Thoroddsen, Lýsing Íslands, I, 213–34.
22
M. Á. Sigurgeirsson, O. Vésteinsson, R . Edvardsson and T. Horsley, Fornleifarannsókn á
Steinboga í Mývatnssveit 2002 [FS189–02072] (Fornleifastofnun Íslands, 2003), pp.8–12;
O. Vésteinsson, ed., Landscapes of Settlement 2002 [FS218–02261] (Fornleifastofnun Íslands,
2004), pp.50–55.
Page 72
Ordernr. 061297 P1
Figure 3.4. Earthworks in Reykjahverfi and Hvammsheiði.
Page 73
Ordernr. 061297 P1
In Skarðagerði (Fig. 3.3) it seems that the third boundary, which is now very
unclear on the surface and could possibly be nothing more than a small enclosure,
was added after a pasture boundary had been built up against the outer perimeter.
In Íradalur it seems obvious that both the boundaries were built later than a
tremendous earthwork that stretches all the way across Hvammsheiði heath.
The most bizarre case of a double boundary enclosure is, however, at Hrísakot,
where both boundaries are almost exactly round and stand extremely close
together leaving, except on one side, a strip of only about 5 m wide between them.
Another unusual case, of a different kind, is a previously unknown site just outside
the homefield of Saltvík. This is the largest ruin area of all and holds at least five
or six longhouse-like structures and a number of smaller structures. This site has
absolutely no boundaries at all — possibly pointing to a very short-lived settle-
ment, or even a function that was not primarily agricultural. However, a small
excavation took place there in 2003 and excluded the idea of a short-lived or failed
settlement, as at least two of the structures have quite thick occupation layers and
were in use after c.950 — demonstrated by the turf in the walls having traces of a
tephra from that time.
Although there is little direct evidence for the age of the sites discussed above,
some of the structures have the characteristics of longhouses — the typical Viking
Age dwelling of the type that is normally dated to between the ninth and twelfth
centuries. The size range of the structures is normally between 18 and 24 m in
length and 6 and 8 m in width. Many of the sites have obviously been reused, and
the most distinct structures probably date to the nineteenth or even early
twentieth centuries. Sites such as Litla-Saltvík and Stekkur have had so many
phases of reuse that it is impossible to see which structures are original, and in fact
impossible to make any suggestions about the nature of the sites.
What is most striking about all these ruins is not only the excellent preserva-
tion of the landscape and the opportunity it affords to study the remains of this
medieval community, but the previous density of the settlement and the large
scale of abandonment. This curious pattern can be highlighted by looking at the
names of surviving farms in the region and comparing them to those of neigh-
bouring areas.
Farm Names in the Study Area
In Iceland it is easy to establish that most of the earliest farms or mother set-
tlements had simple natural feature names. Some examples have already been
mentioned, such as Skörð, Mýri and Reykir. While it is difficult to prove that
Page 74
Ordernr. 061297 P1
secondary farms were not named in this way, it is generally accepted in Iceland —
and makes perfect sense — that farms with suffixes such as -gerði (‘enclosed area’),
-staðir, and -kot (‘cottage’) are later and generally of lower status.23 There are ex-
ceptions; for example, primary farms that were abandoned for some reason often
acquired the suffix -kot. In this way Hóll can become Hólkot and Holt can turn
into Holtakot. If we study the proportion of settlement names in the area (a total
of about thirty) we find a peculiar pattern.
Table 3.1. Settlement names in Reykjahverfi.
Name type No. Names Proportion
natural feature names 6 Reykir, Skógar, Þverá, Skörð, Mýri, Saltvík 20%
Secondary names:
Litla- 4 Litla-Saltvík, Litlu-Reykir, Litlu-Skörð, Litla- 13%
Þverá
-gerði 7 Saltvíkurgerði, Maríugerði (x2), Mýrargerði, 23%
Flatagerði, Skarðagerði, Höfðagerði
-kot 5 Kotið, Hrísakot, Dýjakot, Holtakot, 17%
Brekknakot
-staðir 3 Víkingsstaðir, Uxastaðir, Kárastaðir 10%
other 5 Skálholt, Heiðarbót, Stekkur, Mýrarsel, 17%
Borgarhús
This is not so different from many areas in Iceland, though perhaps the propor-
tion of secondary names is a little higher than usual. However, by 1712 — and
probably much earlier — all the settlements with secondary names had been
abandoned. Although some of the smaller farms, such as Einarsstaðir, Dýjakot and
Holtakot were resettled in the eighteenth and nineteenth centuries, they had been
abandoned for a long time before that according to the land register of 1712. In
most cases their homefields are said to be of no use any more because they are
‘covered in heather and shrub’, or ‘sunken into wetlands’.24 It is highly unusual
anywhere in the country to have almost no secondary farms surviving; in many
districts they account for over half of the total number of farms.
A brief look at the land register for neighbouring areas, such as Aðaldalur and
Reykjadalur, reveals that many secondary farms have survived throughout the ages
— a pattern common throughout Iceland. In both valleys we find quite numerous
-staðir names such as Halldórsstaðir, Öndólfsstaðir, Einarsstaðir; and original
clusters, such as in Hvammar — Presthvammur (prest = ‘priest’s’), Miðhvammur
23
Ó. Lárusson, ‘Úr byggðarsögu Íslands’, in Byggð og saga (Reykjavík, 1944), pp.9–58.
24
Jarðabók Árna Magnússonar og Páls Vídalíns, XI, 245–54.
Page 75
Ordernr. 061297 P1
(mið = ‘middle’), Ystihvammur (ysti = ‘outest’). There are examples of abandon-

ment, for example in the valley of Þegjandadalur, south of Grenjaðarstaðir, where
at least six farmsteads were left in this way. That area awaits further research, but
it remains quite different from Reykjahverfi where the abandoned farmsteads are
spread evenly among the still inhabited farms.
Shielings
There are further results from the archaeological survey in Reykjahverfi — quite
apart from the farm ruins — that are striking. Shielings are almost completely
absent in the area. Only three of the farms had shielings that are known today:
Laxamýri, Skörð and Reykir. All of them have a peculiar location. Skarðasel (sel
= ‘shieling’) and Reykjasel are located unusually far away from the farms, high up
in the mountains to the east, some 10 to 12 km away. Mýrarsel, on the other hand,
was located within easy reach of Laxamýri, just across Lake Mýrarvatn. Results
from archaeological survey in the north of Iceland, especially from Eyjafjörður but
also from Tjörnes, show that both the location and the lack of shielings are very
unusual. In a similar landscape one would expect a shieling on most of the farms.
The expected location in Reykjahverfi, if we use the general results as a compari-
son, would be east and up from the farms, towards the heath and not a great
distance away, perhaps 1 or 2 km. It has to be noted here that as early as 1318
Grenjaðarstaðir had the rights to utilize half of the shieling in Skörð.25
The explanation for this absence of shielings is, of course, not obvious. It is
tempting to suggest that some of the abandoned sites served as shielings, at least
at some point, but yet hard to prove. First of all, the location — with shielings
spread between farms on the best agricultural land — would be highly unusual
and, second, not one place-name supports this. It can possibly be argued that
having shielings in the area served no purpose. Perhaps the land could carry the
weight of grazing and haymaking in the summer; not forgetting the winter grazing
of sheep (especially wethers) practised in Reykjahverfi, at least at the beginning of
the eighteenth century. If this was the case, then there was obviously no point in
taking the animals any further away.
Again, although difficult to prove, especially at this point, it is tempting to link
the lack of shielings to the density of settlement. If a supposed chieftain in the area
was very eager to build up his power, he is likely to have tried to get as many
25
Diplomatarium Islandicum, II, 431.
Page 76
Ordernr. 061297 P1
tenants into his area of influence as possible when organizing the settlement. This
could possibly, although we have no other examples of this, have involved some
drastic limitations, such as using for that purpose all available land that could
possibly sustain a farmstead. Thus no shielings would have been allowed within
the area. This was only meant for farmsteads, with all the remaining land being
used for summer haymaking and winter grazing. It is possible that some of the
shielings were held in common, and used by more than one farm. The ruins of
Skarðasel could support this idea, as they are more extensive than one might
expect at such sites.
This suggestion of population density does not, however, explain the lack of
shielings in later eras, after the settlements were abandoned. There are some indi-
cations from the survey that the management of grazing in summer was different
from many other areas, as an unusual number of structures survive called stekkur
— normally a term for a weening fold used in the spring for separating lambs and
ewes. This could indicate that the ewes, or at least some of them, were kept within
the range of the homeland during the summer and transferred between those folds
depending on the best pasture.
Boundaries – A Brief Overview
Boundaries and earthworks have already been mentioned several times in this
essay, and although we have, at this point, very limited knowledge of them it is
impossible to leave the archaeology of the area without considering them a little
further.26 These structures will presumably become a very important factor in
understanding the settlement landscape in the region. In brief, there is an exten-
sive system of earthworks that stretches over a large area in north-east Iceland —
all the way from Tjörnes in the north, south to Lake Mývatn, and west to Eyja-
fjörður. There are two main types of earthworks in Reykjahverfi (Fig. 3.4): first,
the horizontal ones that stretch from north to south above the farms’ homelands;
and second, transverse ones that lie east–west and thus, at least in some cases,
divide up the space already marked out by the horizontal ones. If we study written
sources that mention the boundaries (mostly place-name records from the nine-
teenth to twentieth centuries), it becomes clear that farmers in the area have
interpreted the horizontal boundaries as one solid earthwork, reaching all the way
26
Á. Einarsson, O. Hansson and O. Vésteinsson, ‘An Extensive System of Medieval
Earthworks in Northeast Iceland’, Archaeologia Islandica, 2 (2002), 61–73.
Page 77
Ordernr. 061297 P1
from Mývatnssveit north to Tjörnes. The transverse dykes, especially the ones that
mark property boundaries, are sometimes mentioned in written sources.
It is possible to speculate how the earthwork system was organized and which
features came first simply by looking at the relationship between individual earth-
works. It seems, for instance, that a boundary that marks the homeland of Skörð
has been built earlier than the pasture boundary to the north of it. Today there is
absolutely no sign of a smaller homefield enclosure at the farm. This boundary
stretches to the south, possibly all the way to the homefield of Heiðarbót. This
could indicate that the farms within it were seen as some kind of unit. The fact
that this enclosure does not include the two extensive ruin areas, Hrísakot and
Skarðagerði, could be an indication that they were built earlier and were clearly
defined properties, marked to the north by the boundary earthwork between
Laxamýri and Skörð. Perhaps this is also an indication of lower status. Further-
more, the boundary markers, both between Laxamýri/Skörð and Reykir/Þverá,
seem to have been built earlier than any of the pasture boundaries linked to them.
This underlines the suggestion made previously that Þverá was settled very early on,
probably as early as the other three farms and, moreover, that markers between farms
were seen as more important than pasture boundaries. Some other boundaries have
quite obviously been built later than the horizontal earthworks and seem to have
fenced off the homelands of the secondary farms: for instance the boundary between
Saltvík/Laxamýri and another one, possibly between Laxamýri and Flatagerði.
Around Saltvík the system gets very complex and needs to be checked in the field
as up till now mapping has mostly been based upon aerial photography.
A few boundaries that seem to belong to the same system have been trenched
in Tjörnes, just north-east of Reykjahverfi, and had all clearly fallen out of use
long before 1477. An earthwork trenched in Fljótsheiði heath, west of Aðaldalur,
was abandoned before 1158. Although the boundary system is most extensive in
this part of the country, earthworks are far from unknown in other parts and some
have even been dated. A few examples in the south show disuse before 1226, but
other parts of the country remain largely unexplored.
Overview and Conclusion
There are still many uncertainties about how settlement in Reykjahverfi took
place and developed. Only limited archaeological research has been done so far,
and few dates for sites and boundaries are available. Therefore it is impossible to
conclude that all sites were settled and/or abandoned at the same time. The pos-
Page 78
Ordernr. 061297 P1
sibility cannot be excluded that some sites are not secondary but primary, original
settlements that failed and were later transferred to more favourable locations.
Despite this, it is clear that the settlement pattern glimpsed through the superbly
preserved archaeology is both exceptional and highly intriguing. Currently some
archaeological work is being carried out in the area. The most extensive project
aims to map all boundaries in a large area in the north-east, including Reykja-
hverfi, using different scales of aerial photographs. This will, without doubt, help
to throw light upon the archaeological pattern preserved. The mapping will be
followed by some limited excavations, providing some dates that will be invaluable
in terms of the boundary and site chronology.
At this point we can speculate a little about the organization of the settlement.
It has already been suggested that a supposed chieftain, most likely in Reykir, was
using a very conscious strategy to build up his power and preserve it against other
chiefs in neighbouring areas, primarily Grenjaðarstaðir and Húsavík. The main
arguments for this are that Reykir had very good quality land; it was the only farm
that possibly had an independent parish church; and the name of the area, Reykja-
hverfi, is significant. It is even possible that the -hverfi name was carefully selected
as it indicates a cluster and thus a strong unit. If, as has been suggested, Reykja-
hverfi did in fact reach further south than it does today, Reykir would have had
a very central location in the district. It can, additionally, be pointed out that
there are quite a few examples in Iceland of geothermal farms becoming centres
of power. All this can be seen as an indication of a community, perhaps a rather
weak one, consciously trying to define itself against other neighbouring commu-
nities that later proved to be superior.
The abandonment of farms is more difficult to comprehend at this stage. It is not

impossible that the community collapsed from within because the area was too
heavily populated and could not sustain such a large number of farms. However,
the fact that almost without exception secondary farms were abandoned raises
questions about organized desertion. This may have been planned by a powerful
person within the area, but it is more likely that outsiders — perhaps in Aðaldalur
— would have benefited from the change, gaining, for example, more area for
upland grazing.
Page 79
Ordernr. 061297 P1
Page 80
Ordernr. 061297 P1
G EOGRAPHY, C OMMUNITIES AND S OCIO -P OLITICAL

O RGANIZATION IN M EDIEVAL N ORTHERN ICELAND
Chris Callow
Introduction
G
uðmundar saga dýra is a narrative covering events in part of northern
Iceland over the course of about fifteen years, c.1186–c.1200.1 It
amounts to fifty-two pages in the standard modern edition of the text
which gives some impression of the level of detail. The saga is anonymous but was
probably written some time in the early thirteenth century. Its author was there-
fore writing about a relatively recent past.2 We have no way of checking the truth
of its stories nor even, for the most part, of providing any sort of alternative ac-
count because none survives, if they ever existed. As a result, any attempt to
reconstruct the past and analyse it based on Guðmundar saga dýra is highly con-
jectural. What follows is an attempt to read the saga with a view to presenting a
picture of what communities consisted of and how they operated in the districts
it covers.
Most of the saga’s action takes place in the Eyjafjörður valley system, mostly in
fact in the smaller valleys immediately to Eyjafjörður’s east and west, and especially
in Hörgárdalur (Fig. 4.1). Guðmundar saga dýra will not be examined in a vacuum,
however. Alongside it we need to consider other sources of information on how
communities might have been defined in this area and at this time: sources such
as legal texts, other sagas, and later ecclesiastical documents and tax registers.
1
Sturlunga Saga, ed. J. Jóhannesson, M. Finnbogason and K. Eldjárn, 2 vols (Reykjavík,
1946), I, 160–212.
2
V. Ólason and others, eds, Íslensk Bókmenntasaga, 2 vols (Reykjavík, 1992–93), I, 316–18.
Page 81
Ordernr. 061297 P1
66 Chris Callow
Figure 4.1. The Eyjafjörður region, where most of the action

of Guðmundar saga dýra takes place.
Page 82
Ordernr. 061297 P1
GEOGRAPHY, COMMUNITIES AND SOCIO-POLITICAL ORGANIZATION 67
The analysis that follows is designed to steer a course between two broadly
separate historiographical traditions. Icelandic scholarship has dealt with the spe-
cifics of institutions and socio-political change over time, while 1980s and 1990s
anthropologically influenced English-language scholarship has concentrated on
examining social processes. The former has generally suffered from being obsessed
with notions of ‘fact’ (notwithstanding literary scholars’ attempts to dismiss sagas
as works of literary fiction),3 while the latter has all but ignored the wide range of
later source material which might flesh out its view of social process. In neither
case has there been a real attempt to identify how communities operated in rela-
tion to geographical space. There are good reasons for this: it is sometimes diffi-
cult to see what is going on beyond the actions described in narratives while, at the
same time, it might be argued that the answers to any investigation into how
communities functioned in medieval Iceland are self-evident to the scholars of
that period. The reader will have to be the judge of whether this study provides
real insights. What follows is an attempt to explore ways of thinking about what
we mean by community in relation to medieval Iceland. The problems of using
the word community have long been appreciated by historians and archaeologists;
the discussion below is not couched in overt language, but I acknowledge the
problems of historians’ use of appropriate language to describe the behaviour of
people in the past as well as the difficulties of imposing overly simplistic models
of community definition.4 Potentially, other approaches, such as social network
analysis, might also offer ways into analysing using the same narrative sources for
early Icelandic society just as they might for medieval societies in general.5
People
Before trying to define what communities existed in medieval Iceland and how
communities were or were not defined by the landscape, a brief assessment needs
3
Ú. Bragason, ‘The Word and the World: A Resonant Textual Fragment of Íslendinga Saga’,
paper given at Sagas and Societies, International Conference at Borgarnes, Iceland, 5–9 September
2002 <http://w210.ub.uni-tuebingen.de/dbt/volltexte/2004/1081/pdf/28_ulf~1.pdf>, takes
a better approach to the issues of literariness, historicity and compiler’s roles.
4
J. Yaeger and M. A. Canuto, ‘Introducing an Archaeology of Communities’, in The
Archaeology of Communities: A New World Perspective, ed. M. A. Canuto and J. Yaeger (London
and New York, 2000), pp.1–15.
5
I am grateful to Roger White for bringing social network theory to my attention. See, for
example, J. Scott, Social Network Analysis: A Handbook, 2nd edn (London, 2000).
Page 83
Ordernr. 061297 P1
68 Chris Callow
to be made of the size of the population for the period under discussion. By doing
this we can begin to understand the possible nature of social, economic and
political relationships between people.
For Iceland in the twelfth century, it is extremely difficult to gauge the size of
the population but it has been a concern of scholars for some time, making use of
later written evidence. While medieval manorial records or their equivalent are
scant for Iceland, the country had a remarkably early census, taken in 1703. The
1703 census and a slightly later tax register (Jarðabók Árna Magnússonar og Páls
Vídalíns)6 provide data that can be measured against occasional references in
earlier literary sources, which give an idea of what was considered a believable size
for some households by certain writers at certain times and in certain places. It is
likely that individual households in the Middle Ages varied a great deal in size, but
perhaps six to eight people per household was typical.7 It has been assumed that
the number of farms in Iceland remained fairly stable from within a century or
two of its colonization to the early modern period, although the evidence for this
is limited. The best evidence is the continuity of place-names and the continued
survival of the largest farms that appear in the medieval narratives. Yet if the 1703
census is anything to go by, in the late twelfth century the population of
Hörgárdalur, the valley in which the major characters of Guðmundr saga dýra
lived, might have been about five hundred.8
Power
Medieval Iceland is often thought of as having a relatively flat social hierarchy, yet
Icelandic society in the late twelfth century was not characterized by an equality
of access to land and therefore did have a clear social hierarchy.9 For the area
under discussion, Guðmundar saga dýra makes it clear that there was a small group
6
Jarðabók Árna Magnússonar og Páls Vídalíns, 13 vols (Copenhagen, 1913–90).
7
Compare A. S. Arnórsdóttir and H. Þorláksson, ‘Saga heimilis á miðöldum’, in Íslenska
söguþingið 28.–31. maí 1997: Ráðstefnurit, 2 vols (Reykjavík, 1998), I, 31–56.
8
Manntal á Íslandi árið 1703 tekið að tilhlutan Árna Magnússonar og Páls Vídalíns
(Reykjavík, 1924–47).
9
For this pattern being established during the initial colonization, see O. Vésteinsson,
‘Patterns of Settlement in Iceland: A Study in Prehistory’, Saga-Book of the Viking Society, 25
(1998), 1–29.
Page 84
Ordernr. 061297 P1
of goðorðsmenn* (‘chieftains’) or höfðing jar* (‘leaders’)10 who lived there and

operated across the region.11 These men and their kinsmen lived on some of the
largest and wealthiest farms in the region (which usually had churches and hence
an income from tithes), and so they had much more of a social advantage than
simply better access to law,12 as some scholars have argued.
Below this top tier it is possible to see — in Guðmundar saga dýra, just as in
other contemporary sagas — a much larger group of relatively well-off farmers
who regularly feature in regional-level politics. Although it is sometimes hard to
see what property these men owned, they probably owned their own farms (which
generally had churches) as well as other farms which they rented out, just as
chieftains seem to have done. This upper tier of society also provided bishops for
Iceland’s two sees — the northern diocese had its seat at Hólar in Skagafjörður —
and a significant number of deacons and priests.13 At the bottom of the social
ladder were the mass of tenants, as well as the servants and vagrants14 who rarely
feature in the narratives except as foot soldiers or spies in particular conflicts.
These people (usually men) are often mentioned by name — often with nick-
names rather than patronymics — but are not said to have lived at any particular
farm. This is probably indicative, on the one hand, of saga authors’ familiarity
with these people but, on the other, of their landless or itinerant status.
Any discussion of communities in medieval Iceland, then, has to appreciate
that while the upper echelons of society were familiar with many of the heads of
10
The word chieftain is used throughout to refer to people identified or identifiable as either
goðorðsmenn (sing. goðorðsmaðr), goðar (sing. goði) or höfðingjar (sing. höfðingi). Goðar is the legal
term for a chieftain who, among other things, acted as a legal representative for his þingmenn (see
note 15 below).
11
Jón Viðar Sigurðsson has suggested that Guðmundar saga dýra is really about control of just
the immediate vicinity of the leading men’s abodes, but the dispute is also about regional control
as we shall see below. G. Karlsson, Goðamenning: Staða og áhrif goðorðsmanna í þjóðveldi
Íslendinga (Reykjavík, 2004), p.155.
12
For access to law, see W. I. Miller, Bloodtaking and Peacemaking (Chicago, 1990),
pp.221–47 (pp.240–47).
13
O. Vésteinsson, The Christianization of Iceland: Priests, Power, and Social Change
1000–1300 (Oxford, 2000), pp.201–09.
14
See, for example, Grágás. Elzta lögbók Íslendinga. Útgefin eptir skinbókinni í bókasafni
konungs, 2 vols (Copenhagen, 1852; repr., Odense, 1974), Ib, 3–28. For an English translation of
this text, see Laws of Early Iceland: Grágás, the Codex Regius of Grágás, with Material from Other
Manuscripts, trans. A. Dennis, P. Foote and R . Perkins, 2 vols (Winnipeg, 2000), II, 29–52.
Vésteinsson, Christianization of Iceland, p.81.
Page 85
Ordernr. 061297 P1
70 Chris Callow
households at other farms, as well as their own retinues containing kinsmen,

þingmenn15 and dependants of various kinds,16 wealth was very unevenly dis-
tributed. Icelandic social structures were probably not so different from those of
much of western Europe. The key differences are, first, that Iceland did not have
the wealth to support an aristocracy that was divorced from the everyday realities
of farming; chieftains were perhaps like poor relatives of the castellans of tenth-
century western Europe, such as the well-known Gerald of Aurillac.17 Second, the
source materials are different. Saga narratives provide the detail of everyday dis-
pute in a form which is lacking for many areas of western Europe, where disputes
are recorded in shorter, more formal documents.18
The Course of Events
A potted account of Guðmundar saga dýra should serve to show how, according
to the saga’s author, power worked in the late twelfth century. The basic frame-
work of the saga is formed by the political relationships between three men and
their relatives and associates. The geographical dimension of these relationships
needs to be stated first. Guðmundr dýri Þorvaldsson lived at Bakki in the rela-
tively small and sparsely populated valley of Öxnadalur.19 His most obvious rival
was the chieftain Önundr Þorkelsson who lived first at the mouth of the same
valley system at the farm of Laugaland in Hörgárdalur before moving to Langahlíð
in the same valley (ch. 9). After Önundr’s death (ch. 14), his role in local politics
was taken up by his son-in-law Þorgrímr alikarl who lived at Möðruvellir in
Hörgárdalur when the saga began. Möðruvellir, little more than 4 km north of
Laugaland as the crow flies, had been the home of another chieftain, Þorvarðr
15
Þingmenn, sing. þingmaðr, were those men who were not goðar but who owned sufficient
wealth to be obliged under Icelandic law to declare themselves as supporters of a goðorðsmaðr. In
practice this was usually at least one person per farm. Miller, Bloodtaking and Peacemaking,
pp.22–25, gives a succinct account of the system.
16
The raising of retinues forms a key part of how communal activity can actually be identified
in the Icelandic narrative sources as will be seen below.
17
Odo of Cluny, Life of Gerald of Aurillac, in Soldiers of Christ, ed. by T. F. X. Noble and
T. Head (London, 1995), pp.296–362.
18
W. Davies and P. Fouracre, eds, The Settlement of Disputes in Early Medieval Europe
(Cambridge, 1986).
19
Fig. 4.1. For the sake of convenience in this article, Öxnadalur will be considered to be part
of Hörgárdalur, from which it branched. The latter term will refer to both valleys.
Page 86
Ordernr. 061297 P1
Þorgeirsson. Þorvarðr’s son, Ögmundr Þorvarðsson, had been in Norway for some
unspecified time and when he returned (ch. 10), rather than living in Hörgár-
dalur, he eventually settled some distance away at Háls in Fnjóskadalur, east of
Eyjafjörður. The saga is not explicit about what had happened for Ögmundr and
his father to lose control of Möðruvellir to Þorvarðr.
The overall pattern of politics in the region centring on Hörgárdalur is one in
which tensions continue to rise between Guðmundr dýri and his main rivals,
Önundr and then his son-in-law Þorgrímr alikarl, until Guðmundr defeats them
each in turn. Each of these men has þingmenn and other supporters across the
valleys (Fig. 4.2). Guðmundr dýri is said to have dispensed with the local assembly,
the Vaðlaþing (Hann hafði af tekit Vaðlaþing. Skyldi þar eigi sóknarþing heita)20
which, if we believe the simplistic, legal texts’ views of local þing*, was a disruption
of an ages-old folk moot. While Guðmundr dýri may not, in fact, have been the
first chieftain to be in a position to suspend a local assembly, the text sees this as
a significant departure for northern Icelandic politics in the later twelfth century
and a measure of Guðmundr’s growing power. Guðmundr dýri and Önundr and
Þorgrímr alikarl grow in power, as demonstrated by the eventual disappearance
of Þorvarðr Þorgeirsson from the narrative and Önundr’s and Þorgrímr alikarl’s
takeover of farms in lower Hörgárdalur. Guðmundr dýri’s considerable power can
also be measured by the fact that when one of his supporters falls out with
Ögmundr Þorvarðsson, the case against Guðmundr’s supporter is led by Jón
Loptsson, the leading chieftain from the south (ch. 10). Guðmundr’s ‘side’ loses
this case but he gets away with not paying the enormous fine awarded against him.
From across the region to the north of Hörgárdalur — Fljót, Ólafsfjörður and
Svarfaðardalur — people still seek out either Guðmundr or Önundr as their
advocate (chs 11–14). Guðmundr dýri still also has the support of neighbouring
chieftains against Önundr and, together with Kolbeinn Tumason, Guðmundr
carries out the so-called Önundarbrenna, the arson attack on Önundr’s home at
Langahlíð which kills Önundr and many others (ch. 14). Guðmundr now buys off
Ögmundr Þorvarðsson by belatedly paying off the 1200 hundruð* fine which he
owed from their previous dispute, and thereby ending any chance of Ögmundr
acting against him.
For a brief while peace seems to have broken out, but Þorgrímr alikarl and
Önundr’s sons then make an attempt to remove Guðmundr from the political
scene. As they gather troops for a raid on Guðmundr at Bakki, they kill a
20
Sturlunga Saga, ed. Jóhannesson, Finnbogason and Eldjárn, I, 170; J. V. Sigurðsson, Frá
Goðorðum til ríkja (Reykjavík, 1989), p.58.
Page 87
Ordernr. 061297 P1
72 Chris Callow
Figure 4.2. Map to show Hörgárdalur and surrounding valleys.
supporter of Guðmundr’s who lived at Arnarnes on the west coast of Eyjafjörður

(chs 17–18). They raid Bakki, but Guðmundr is not present and instead they only
remove his men’s ‘weapons and shields’. At this point the bishop of Hólar visits
Hörgárdalur to encourage each side to back down. Þorgrímr alikarl goes away to
the south of Iceland to gather support from his kinsmen there which allows
Guðmundr (with Kolbeinn Tumason again) to raid Laugaland and other farms
in lower Hörgárdalur.
A further attempt by neutrals to negotiate a long-term peace ends in failure,
and Þorgrímr alikarl leads a raid on Bakki with a force of one hundred men (ch.
Page 88
Ordernr. 061297 P1
23). This raid ultimately fails, however, and leads to Þorgrímr’s force being
surrounded at a neighbouring farm by men led by Guðmundr, Kolbeinn Tumason
(who is said to have brought six hundred men) and Ögmundr Þorvarðsson. Þor-
grímr leaves Hörgárdalur and is said to have moved to the farm of Geldingaholt
in Skagafjörður. The saga ends with three episodes not related to the longer-term
conflict mentioned above or to each other. The latter two are worth mentioning
here. The farm of Vellir in Svarfaðardalur is a source of dispute for Guðmundr’s
supporters the Arnþrúðarsons, whose father had occupied the farm (ch. 25). As
it is a church farm (staðr), however, the bishop of Hólar tries to appoint his own
tenant rather then one of the Arnþrúðarsons. In the end a compromise is reached
whereby the farm will be run by the bishop’s tenant but revert to the Arnþrúðar-
sons after his death; Guðmundr is noticeably cautious in his opposition to the
bishop. The saga ends with a story about Guðmundr which the saga author seems
to have not known quite where to place chronologically and so saved for the end.
This is an account of a dispute in which a Norwegian trader cuts off the hand of
a deacon. Eventually, after the Norwegian refuses to pay compensation, Guð-
mundr dýri intervenes and offers him the chance to choose between paying or
losing his own hand. The compensation is paid quickly. The penultimate sentence
of the saga says that Guðmundr became a monk at Þingeyrar in Húnavatnssýsla,
west of Skagafjörður.
To recap, then, the saga sees the three groups (Guðmundr dýri; Önundr and
Þorgrímr alikarl; Þorvarðr Þorgeirsson and Ögmundr Þorvarðsson) working as
advocates in disputes between less powerful men. The disputes which the saga
describes are otherwise quite varied in their causes, their personnel, the point at
which the chieftains get involved and the precise nature of their resolution.
Geographically they are also quite well dispersed over the region centring on
Hörgárdalur although there are patterns, to which I shall return. Almost any
episode might therefore serve as examples; two are chosen here. Neither is truly
typical but they are representative in many respects and have the advantage of
being relatively self-contained and involving relatively few actors.
The saga begins with a dispute over a farm in Reykjadalur, east of Eyjafjörður,
called Helgastaðir (chs 1–3). The dispute begins after its owner, Guðmundr, be-
comes a monk, leaving the property to his son. The son dies before him, however,
and so nobody is now quite sure who the heir is. A wealthy neighbour, Eyjólfr
Hallsson, decides to buy the property from the former owner-now-monk at a
knock-down price. Opposing the neighbour, however, are the brothers of the now
monk; that is, the uncles of the deceased owner. In the end the uncles each turn
to their respective chieftains, Önundr Þorkelsson and Þorvarðr Þorgeirsson, who
Page 89
Ordernr. 061297 P1
74 Chris Callow
only agree to prosecute the case on the grounds that they themselves are given
ownership of the farm. Eyjólfr, the neighbour who thinks he has already bought
Helgastaðir, gets support from people from Skagafjörður, far to the west, and from
someone from Eyjafjörður who may also have been a chieftain. Eyjólfr’s advocates
succeed in court in getting Önundr and Þorvarðr made outlaws for supposedly
having stolen Helgastaðir.
Rather than being in fear of their legal status as outlaws, Önundr and Þorvarðr
arm themselves and take fifty men to occupy Helgastaðir. There is a physical
stand-off between Önundr and Þorvarðr’s force and that of Eyjólfr and his sup-
porters. Guðmundr dýri and another goðorðsmaðr from Skagafjörður intervene at
this point, arriving with their own force to prevent any fighting. In the end an
agreement is reached involving Guðmundr dýri and his associate taking on the
role of mediators. It is agreed that Helgastaðir is to be given to the son of Eyjólfr’s
advocate from Eyjafjörður, one Klængr, rather than directly to either of the
original claimants. Rather than this being a compromise, however, it transpires
that Klængr later sold the property to Eyjólfr. Thus Guðmundr dýri was partly
responsible for preventing his powerful neighbouring rivals from acquiring a
valuable property.
The second example of a dispute is one which takes up a single chapter, chap-
ter 8. It describes a dispute between a man from Garðshorn in Svarfaðardalur,21
Brandr, and a man called Sumarliði from the farm of Tjörn, a few kilometres
north of Garðshorn. The dispute began when Brandr went to church and his
horse got loose and ate the hay in the homefield of Tjörn without Sumarliði’s
permission. Sumarliði expressed his disgust by hitting Brandr and his horse with
a cudgel (með lurk) at the scene. At another feast day, but seemingly a considerably
larger social event, at the kirkjudagr22 of Öxnahóll church in Hörgárdalur, Brandr
attacked and killed Sumarliði. Brandr’s escape (to Norway) was aided by two men
from Skagafjörður, namely the chieftain Kolbeinn Tumason and Bjärn Ólafsson
from Fell. Brandr and his associates were prosecuted for the killing by Guðmundr
dýri who, in a show of regional solidarity, was supported by Önundr. A further
minor adjustment was made to the local political scene with the expulsion of one
of Brandr’s accomplices from the north.
21
Fig. 4.1. Sturlunga Saga, ed. Jóhannesson, Finnbogason and Eldjárn, I, 174; Vésteinsson,
this volume.
22
A kirkjudagr was an annual celebration of a church’s consecration. Six such events are
mentioned in the Sturlunga Saga compilation: Sturlunga Saga, ed. Jóhannesson, Finnbogason and
Eldjárn, II, 481.
Page 90
Ordernr. 061297 P1
These two disputes demonstrate some of the themes which will appear below.
They show, for example, how a dispute in one valley could soon escalate to involve
powerful advocates from other areas, whether or not a chieftain lived in the valley
where the original disputants lived. The wide variety of contacts which an indi-
vidual might have in this society, regardless of their wealth or social standing, also
comes out. The geographical range of communication is also quite considerable.
Finally, the two disputes here also demonstrate that, depending on the nature of
the dispute, the same goðorðsmenn could work for or against one another. Alle-
giances and affinities could change according to particular circumstances.
Social Events
It is worth asking when people who lived relatively close to each other actually
met up in late twelfth-century Iceland. For Iceland, relatively dispersed settlement
pattern would have meant that spontaneous socializing would have had to have
been relatively well planned (unless it involved just people on the same farm and
its associated outliers),23 and was much more likely to have taken place in summer.
At the same time, interaction between neighbouring farms was probably not
uncommon especially if they were near and people owned horses, as frequently
seems to have been the case. It was while making a superficially ordinary visit to
Arnarnes, for example, that Hákon Þórðarson from Laufás is said to have killed
his lover’s husband (ch. 5). In this case Hákon is supposed to have visited
Arnarnes — about 7 km across the fjord from Laufás but still arguably a neigh-
bouring farm — throughout the winter. Laufás and Arnarnes were relatively large
farms and one might suspect that the senior men of the household had more free
time to pursue their love interests than many other people, but visits between
farms were probably quite regular. Þorvaldr from Ytri-Bægisá and a húskarl made
shorter journeys for love across Hörgárdalur, to visit a mother and daughter both
called Birna who lived at Efri-Langahlíð (ch. 12). Similarly, people were likely to
meet when out shepherding or when working close to farm boundaries. The
annual hiring or ‘moving days’ (fardagar) in May also circulated people and
information.24
23
Vésteinsson, this volume.
24
Grágás, IA, 128–31.
Page 91
Ordernr. 061297 P1
76 Chris Callow
More regular events and patterns of behaviour existed, however. Going to

church25 and taking part in wedding celebrations are two examples which are
much more likely to have seen large groups of people get together during the
year.26 Some sort of communal sheep round-up in autumn was also likely to have
been important.27 As sagas are essentially records of things which happen once or
are ‘eventful’, then we ought to expect them to record quite a lot of these kinds of
events. Indeed, in the contemporary sagas such events do get recorded and they
do show large numbers of people socializing, at least at the upper end of the social
ladder. Weddings are mentioned in Guðmundar saga dýra (e.g. ch. 5) but it is
difficult to assess how many people attend and where the guests are from, just as
is the case in other Contemporary Sagas.28 Sagas of Icelanders are less reticent
about numbers but they clearly generalize and exaggerate. Nevertheless, relatively
short life expectancy, probably even by medieval western European standards, and
a cultural preference for girls to marry early may have led to a comparatively high
number of weddings in medieval Iceland.
Church attendance might not have been particularly regular in twelfth-
century Iceland but it is clear that many churches which can be documented in
the later Middle Ages existed by this stage. It can be inferred that others existed
too.29 Two things need to be said about churchgoing: first, that it happened in
Guðmundar saga dýra, and a system of parishes* is likely to have existed and to
have shaped community identity; second, that there were religious festivals that
had a wider importance and involved the attendance of a wider community.
Although our understanding of the evolution of parishes is hampered by a lack
of early documentary evidence for northern Iceland, it would seem likely that
parishes were in existence well before they were first systematically recorded for
25
O. Vésteinsson, ‘Íslenska sóknaskipulagið og samband heimila á miðöldum’, in Íslenska
söguþingið 28.–31. maí 1997, I, 147–66 (p.161).
26
For a description and analysis of a similar situation, see W. Davies, Small Worlds: The
Village Community in Early Medieval Brittany (London, 1988), especially chapter 5.
27
See Grágás, Ib, 113–21, for the regulation of pasture.
28
Contemporary Sagas (samtíðarsögur) is the collective name for texts written in the thir-
teenth or fourteenth centuries which describe events in Iceland in the relatively recent past. This
genre can be sub-divided into two groups: the compilation of texts known as Sturlunga saga,
named after the family with which they are associated, and sagas about individual bishops
(biskupasögur).
29
Vésteinsson, Christianization of Iceland, esp. pp.37–45, 94–101 and Tables 1 and 5.
Page 92
Ordernr. 061297 P1
the diocese of Hólar in 1318.30 It is also true that it is not always possible to
identify all of the farms which made up the early parishes because in the earliest
charters identifying church property (máldagar, sing. máldagi) individual farms
are not named. Not all churches have equally old charters either and so compari-
sons can be difficult. This is especially relevant to Hörgárdalur where máldagar
only exist for Bakki, Bægisá, Laugaland and Auðbrekka from a document dated
to 1394.31 Möðruvellir became a monastery from 1296 and, as this church is
omitted from the 1318 compilation, no máldagi survives for it at all as a simple
church. Nevertheless the size of parishes or at least the number of farms from
which they took dues is instructive for gauging something of the nature of the
church’s administration and, by proxy, some sort of social organization.
For Hörgárdalur, the máldagar are not particularly forthcoming but in 1394
Bakki received lighting and hay dues from ten farms. Auðbrekka, the only other
church in the valley for which there is a clear statement on the matter, received
hay dues from eight farms but, unusually, received lighting dues from a smaller
number, only two. For Eyjafjörður and its tributary valleys it would seem that it
was usual in 1318 for most churches to receive dues from between six and twelve
farms. Given the number of farms recorded in later centuries and the evidence we
have for churches in Guðmundar saga dýra, it would seem possible that the pat-
tern was the same in the late twelfth century. Indeed, the saga’s evidence would
suggest so as attendance at the parish church provided the opportunity for dis-
putes between people who lived close to one another, such as when Brandr and
Sumarliði clashed outside the church at Tjörn (ch. 8). Unusual events, such as the
kirkjudagr at Öxnahóll (ch. 8), brought in people from further around, but on this
particular occasion, the only evidence we have is that those from outside the
parish were mostly well-off; Brandr, the son of a farmer on a relatively modest
farm (Tjarnargarðshorn), may only have been present because he knew that the
better-off Sumarliði would be there.
Other events and places also warrant brief discussion. Horse fights and games
do take place in this saga just as they do for Sagas of Icelanders. A horse fight in
the Fljót district is the scene for an attack on Nikulás Skratta-Bjarnarson, a
þingmaðr of Guðmundr dýri who lived at Grindill in the same district (ch. 12).
Shortly afterwards, at Syðri-Bægisá, a farm close to the confluence of Hörgárdalur
and Öxnadalur, there were games (ch. 12). The events recorded for this show a
30
Reykjavík, 1853–1976), II.
31
Diplomatarium Islandicum, III.
Page 93
Ordernr. 061297 P1
78 Chris Callow
‘coming together’ of people from across the valley and beyond as tensions rise
between a supporter of Kolbeinn Tumason’s (a kinsman of the Nikúlas who had
just been attacked) and men from Langahlíð, now the home of the chieftain
Önundr. Taken at face value, the saga seems to imply that contentious social
events were organized one close after the other. It is more likely, however, that
although such occasions were held regularly they were less memorable because
they passed off peaceably, or else any conflicts did not involve local leaders so
directly. Another measure of their regularity and importance is also suggested
when the saga shows the abnormality of Guðmundr and ‘his men’ not going to
meetings and games for several months as local political tensions rose.32
Markets and assemblies figure fully as well, and it is clear that these sorts of
business-orientated gatherings of more people were important. Gásir, the port on
the west coast of Eyjafjörður close to the mouth of Hörgárdalur, formed the
backdrop to a handful of events in Guðmundar saga dýra and other narratives.
Courtship, or an arranged marriage, between Guðmundr dýri’s daughter Ingibjörg
and Þorfinnr Önundarson takes place at Gásir (ch. 9). It is possible that Gásir
provided some sort of neutral venue for local chieftains to meet.
Secular Institutions
The historiography and realities of early Icelandic society demand that attention
be paid to goðorðsmaðr-þingmaðr relationships and communes, two forms of social
relationship which were unavoidable according to the legal texts.
Men owning a certain amount of property were obliged to ally themselves
publicly to one or other of the men who inherited a goðorð (chieftaincy), who, as
we have already seen, were known as goðorðsmenn or goðar. In practice it would
seem from all forms of evidence that most men who were heads of household, as
well as others, were likely to have owned the wealth which made them eligible as
þingmenn. Whether, in actuality, everyone was aware at all times of such relation-
ships is another matter. Certainly, it was only the wealthier farmers who seem to
have made use of their goðorðsmenn when they needed them to act as advocate or
protector. Where those men identified as þingmenn in the saga can be associated
with a farm, the farm is one of middling to high value in Jarðabók Árna Mag-
nússonar og Páls Vídalíns or other early modern tax registers.
32
Sturlunga Saga, ed. Jóhannesson, Finnbogason and Eldjárn, I, 185: ‘Þá fór Guðmundr aldri
til mannamóta um sumarit ok engir hans menn ok eigi til leika um vetrinn.’
Page 94
Ordernr. 061297 P1
Other people could actively ally themselves to a chieftain but, rather than
being given the title þingmaðr in Guðmundar saga dýra, are identified as fylgðar-
maðr, ‘follower’ or ‘supporter’. It appears that for the author of Guðmundar saga
dýra, þingmaðr and fylgðarmaðr were not synonymous. In the three places where
fylgðarmaðr is used, all qualify the term with another: Björn Gestsson from Sandr
was a fylgðarmaðr and þingmaðr of Önundr’s (ch. 4) while Björn Hallsson, his
father and brothers and, in the same chapter, Þórðr Þórarinsson, were all fylgðar-
menn ok vinir (‘supporter and friend’) of Þorvarðr Þorgeirsson’s (ch. 10). In other
words, there would seem to be two different kinds of relationship which the
author could identify as connecting weaker men with a chieftain.
Of equal significance to the fact that some men are identified as þingmenn and
fylgðarmenn is the fact that none of these men lived close to chieftains, that is, in
the valleys in which the chieftains lived. We do not know explicitly about the legal
allegiances of farmers closer to chieftains’ main residences because, perhaps, there
was less freedom for farmers to choose.33 Guðmundar saga dýra’s lack of interest
in the relationships of farmers to goðorðsmenn in Hörgárdalur would strongly
suggest that the author assumed that his audience would know what they were.
This might suggest that the author assumed too much with respect to his audi-
ence’s local knowledge or that, just as likely, it was unquestionable that entire
valleys would be very firmly controlled by one chieftain. The exact way in which
the valley may have been controlled will be discussed below.
The lack of ability to choose a goðorðsmaðr may also have been a force for
group cohesion. In one instance in the saga it would appear that it actually united
local populations against chieftains when they overstepped the mark, as seems to
have happened when Hörgdælir, people from Hörgárdalur, complained to the
bishop of Hólar about Guðmundr dýri (ch. 15).
Communal organization may well have been well defined at a local level even
if our literary evidence does not take much interest in it. Hreppar* (sing. hreppur)
or communes are identified in Grágás, the body of legal material which has been
referred to at varying points throughout this essay. According to Grágás, hreppar
were associations of no fewer than twenty farmers who each were liable for
þingfarakaup, the assembly attendance contribution payable to a goðorðsmaðr.34
33
C. P. Callow, ‘Landscape, Tradition and Power in a Region of Medieval Iceland: Dalir
c.900–c.1262’ (unpublished doctoral thesis, University of Birmingham, 2001), pp.187–96.
34
Kulturhistorisk leksikon for nordisk middelalder, 22 vols (Reykjavík, 1956–78), VII, 18–20;
W. Gerhold, Armut und Armenfürsorge im mittelalterlichen Island, Skandinavistische Arbeiten,
18 (Heidelberg, 2002), pp.136–73.
Page 95
Ordernr. 061297 P1
80 Chris Callow
The farms of these men had to be in a contiguous block,35 making them territorial
units like parishes. The main purpose of the commune was supposedly to ensure
the distribution of the tithe to the poorest members of society, and the control of
the movements of the landless between different localities in order to not over-
burden the commune members who had an obligation to support them. Thrice-
yearly meetings were to ensure that these functions were carried out. The main
problem with the idea of communes, however, is that they occur rarely in the nor-
mative saga material. Only a handful of references to the term exist in surviving
saga literature, one of which is from chapter 25 of Guðmundar saga dýra, and a
couple of others are found in Sagas of Icelanders which cover the same region.36
The example from Guðmundar saga dýra is typical of other references to events
in the text or any other saga, in that it is there because it is the scene of an attack
by one man on another. The so-called hreppstefna (‘hreppur meeting’) takes place
at the farm of Háls in Svarfaðardalur, a relatively valuable farm (sixty hundruð in
the early eighteenth century)37 at the mouth of Svarfaðardalur, which was cen-
trally placed within the hreppur of Svarfaðardalur that can be identified in early
modern sources. Háls is slightly unusual for a farm of its size, however, in appar-
ently not having had a church which survived into the early modern period. In a
document dated to 146138 it is said to have had a chapel* which had fallen into
disrepair, and it appears in a list of the possessions of the monastery of Möðru-
vellir in 1447.39 The fairly central geographical location of Háls within this
hreppur, though, makes sense of its function as a central point or meeting place,
despite the farm’s situation away from the populated inland parts of Svarfaðar-
dalur itself. It would be most reasonable for meetings to have taken place at Háls
when people were travelling in and out of the valley to the coast. Given that the
three annual meetings of hreppar are likely to have preceded or followed meetings
of assemblies of regional and national significance in spring, summer and autumn
(respectively várþing, Alþing* and leið), Háls would have made a sensible stopping-off
point for the hreppur’s population, regardless of its status otherwise.
35
Grágás, Ib, 171.
36
Víga-Glums saga, ch. 13, in Eyfirðinga sögur, ed. J. Kristjánsson, Íslenzk fornrit, 9
(Reykjavík, 1956), p.43. Reykdœla saga, ch. 7, in Ljósvetninga saga, ed. B. Sigfússon, Íslenzk fornrit,
10 (Reykjavík, 1940), p.169.
37
Jarðabók Árna Magnússonar og Páls Vídalins, X , 93–94.
38
Diplomatarium Islandicum, V , 356.
39
Diplomatarium Islandicum, IV , 712.
Page 96
Local Communities
The word community has been used sparingly in the discussion above. Some men-
tion has been made of space or distance, and of a variety of ways in which people
interacted with those living at other farms and the reasons behind those interactions.
This final section, however, is designed to flesh out the picture of communities’
form by analysing Guðmundar saga dýra’s depiction of socio-political interactions
at the level of the individual valley. ‘The valley’ is about as good an analytical unit
for Iceland as there is, although there are obvious problems with it. It must be
accepted — and the earlier discussion should have made this clear — that the
often difficult terrain in Iceland did not prevent fairly frequent communication
between areas tens or hundreds of kilometres apart. Individual valleys were rarely
isolated, at least not in northern central Iceland.
Hörgárdalur provides the best case study for seeing society function in Guð-
mundar saga dýra because this valley system was a home to chieftains. Although,
inevitably, the text does not cover everyday events, the unusual happenings it
purports to record do give a sense of how local society operated. Several farms in this
area are mentioned (beginning in the northernmost, outer part of the valley and
going south and up): Hlaðir, Möðruvellir, Dynhagi, Laugaland, Auðbrekka, Langa-
hlíð,40 Efri-Langahlíð,41 Grund,42 (Efri-?)Rauðalækur, Hallfríðarstaðir, Öxnahóll,
Ytri-Bægisá, Syðri-Bægisá, Myrkárdalur, Steinsstaðir and Bakki (Fig. 4.1). This list
nevertheless demonstrates that we have only a very partial view of the population
that lived in this area. Almost all of these farms were large in the early modern
period; and if the tax registers are anything to go by, we ought to expect there to
have been over sixty households or dwellings in the valleys inland from Hlaðir.43
First something needs to be said about the possible nature of medieval hreppar
in this small valley system. Two hreppar existed here in the early modern period,
Skriðuþingsókn and Glæsibæjarhreppur. The first covered most of the farms listed
above, all bar Laugaland, Rauðalækur and Ytri-Bægisá, and included all of Hörgár-
dalur and Öxnadalur except for the former’s eastern side from Ytri-Bægisá north-
wards. Glæsibæjarhreppur formed a kind of U-shape on the map, beginning in
40
Modern Skriða, Sturlunga Saga, ed. Jóhannesson, Finnbogason and Eldjárn, I, 552.
41
Modern Langahlíð, Sturlunga Saga, ed. Jóhannesson, Finnbogason and Eldjárn, I, 552.
42
Grund is not identified in Fig. 4.1 because it is not identifiable as a modern farm in
Hörgárdalur but probably close to modern Skriða, Sturlunga Saga, ed. Jóhannesson, Finnbogason
and Eldjárn, I, 553.
43
Vésteinsson, this volume.
Page 97
Ordernr. 061297 P1
82 Chris Callow
lower Hörgárdalur and extending round as far as Glerá (including that particular
farm, which was named after the river next to it (Fig. 4.1)). In practice, therefore,
Guðmundar saga dýra’s account of local politics might be seen as only dealing in
detail with one hreppur. The number of farms from Glæsibæjarhreppur which
feature is no more than for many, far more distant hreppar, although it is also true
to say that the little we know of Skriðuþingsókn derives from a handful of
episodes in the saga.
The names of the two hreppar also deserve brief discussion. They are named
after farms: Skriða was the name of Langahlíð from the later Middle Ages, after
a substantial collapse of the slope had destroyed the original farm; Glæsibær was
a church farm of at least the same size as many of the larger farms in Hörgárdalur.
Although it can be argued that the early modern names may only indicate the
significance of the eponymous farms in a later period, the medieval evidence could
imply some sort of earlier primacy for these farms. Langahlíð was a farm which
Önundr sought to move to in preference to Laugaland. The absence of Glæsibær
from the saga is interesting, and it is best explained, if not wholly satisfactorily, as
reflecting the power of Önundr at Laugaland and the text’s disinterest in
Glæsibæjarhreppur generally.
Regardless of the paucity of our information, and perhaps because of it, we can
see something which resembles a very local level socio-political community in
action in Guðmundar saga dýra’s Hörgárdalur. We can see its complexities and
the importance of certain wealthy individuals, the chieftains, in influencing oth-
ers’ lives. The first time the saga mentions anyone from Hörgárdalur other than
a local leader, it identifies two priests who are asked to arbitrate in a dispute
between supporters (from outside Hörgárdalur) of Önundr and Guðmundr
(ch. 4). The priests are from the farms of Möðruvellir and Öxnahóll respectively.
As in many disputes in saga literature, it seems likely that these men were chosen
as arbitrators because they were each regarded as loosely allied to one of the
disputants but could be relied upon to be more moderate than the disputants
themselves. As a priest at Möðruvellir, Hallr Gunnarsson was in the household of
Þorvarðr Þorgeirsson who had worked with his fellow chieftain, Önundr. Björn
Steinmóðarson, on the other hand, had married Guðmundr dýri’s daughter but
lived at Öxnahóll. The human geography here is as important as the personal ties
because each arbitrator lived at a large farm in the same hreppur as one of the
disputants, but each lived closer to the chieftain with whom they had other
connections than to the other one.
As a general rule it would seem that Skriðuþingsókn, if it had the same form
in the medieval period, was divided on roughly geographical lines. This does not
Page 98
Ordernr. 061297 P1
mean to say that everyone in a single household or all farms in a ‘sub-community’44

shared strong allegiances to a single chieftain at all times. Nevertheless, Guð-
mundr dýri had more support and more land in the upper parts of the valley(s),
and Önundr and Þorgrímr alikarl had more support in the lower valley. Þorvarðr
and his son Ögmundr are noticeably absent from Hörgárdalur politics in the latter
half of the saga and their interests were much more obviously focussed on Fnjó-
skadalur. Thus Guðmundr was able to house a concubine whom he had brought
in from Svarfaðardalur at Myrkárdalur, just over the ridge from Bakki (ch. 9). He
also owned Bakki’s neighbouring farm, Steinsstaðir (ch. 6). Efri-Langahlíð, too,
is a property which Guðmundr indirectly has an influence over, after his daughter
divorced the aformentioned Björn Steinmóðarson and she was given the farm as
part of the divorce settlement (ch. 12).
Önundr and Þorgrímr alikarl and their supporters are also fairly easy to
identify. Nearer the coast, they control Laugaland, Langahlíð and Möðruvellir,
and have supporters at Hlaðir, Rauðalækur and Dynhagi. Björn Steinmóðarson
still lived at Öxnahóll and his allegiance is definitely with Önundr after his divorce
(ch. 13), something for which he suffers public humiliation at Guðmundr’s hands
and which suggests the strength of his falling-out with Guðmundr (ch. 19).
Björn’s farm was halfway up the valley, close to where Öxnadalur and Hörgárdalur
fork, and must have served as an important meeting point and lookout post at
various times. Björn’s allegiance was of vital importance to the chieftains, espe-
cially as he probably had enough wealth and a large enough household to resist the
more extreme forms of bullying which other farmers received.
The allegiance of Þórir Bárðarson, the farmer from Langahlíð,45 whom
Önundr forcibly ejects and who moves to Laugaland (ch. 9), is not clear-cut.
Although it rather contradicts the idea that Langahlíð was an early centre of
power, the text does at least point out that there was a considerable difference in
Önundr’s and Þórir’s standing (ríkismunr).
The chequered history of a man called Illugi helps confirm the general pattern.
Illugi gained the support of two chieftains but ultimately settled at Hlaðir, next
to Möðruvellir, where Þorgrímr alikarl lived (ch. 9). When Illugi is first men-
tioned, he is said to have killed someone, but he stays with Guðmundr dýri at
Bakki and is aided by Guðmundr’s allies and kinsmen in Svarfaðardalur. After a
44
Of the kind where a number of dwellings were clustered together, such as at Garðshorn or
Langahlíð but often assumed to be a single ‘farm’. Vésteinsson, this volume.
45
Sturlunga Saga, ed. Jóhannesson, Finnbogason and Eldjárn, I, 176.
Page 99
Ordernr. 061297 P1
84 Chris Callow
year away from Iceland, no doubt enforced by Guðmundr, Illugi comes back to
spend a season at Möðruvellir, now run by Þorgrímr alikarl, before eventually
marrying and settling at Hlaðir. Illugi may have consciously changed allegiances
with the result that he was himself murdered by someone trying to gain support
from Guðmundr (ch. 16).
Other individuals or farms seem to have been able to stay relatively neutral or
else win the respect of the opposed chieftains; here there are signs of an identi-
fiable community. When Guðmundr comes to kill Önundr (ch. 14), a man called
Gálmr from the farm of Dynhagi, close to Möðruvellir and Laugaland, was with
Önundr. Yet he is described by the saga as also being a friend of Guðmundr and
Kolbeinn Tumason, and so the latter pair offered him the chance to escape. Gálmr
chose to stay and presumably died in the fire.46 Tjörvi from Rauðalækur, a farm
midway between Bakki and Laugaland, is also granted a truce by Guðmundr’s
party but dies of his wounds.
Occasional clues also show that connections between households were plenti-
ful and that territorial divisions should not be exaggerated. Kálfr Guttormsson,
for example, who was Guðmundr’s kinsman by marriage, lived at Auðbrekka and
was targeted by Þorgrímr alikarl’s party as part of Þorgrímr’s attempt to avenge
Önundr. Þorgrímr’s party seems to have been singling out Guðmundr’s kin living
in lower Hörgárdalur and, having just failed to attack a kinsman of Guðmundr’s
at Langahlíð, made for Kálfr. Kálfr was saved from death, however, because of his
friendship with one of the attackers (ch. 23).
Regional Communities
The social relationships or collective activities detailed above must have been what
community was about for most of the population in medieval Iceland most of the
time. It is also important to note, if only briefly here, that supralocal communities
of landowners — the goðorðsmenn and höfðingjar — were not insignificant. The
geographical dimensions of these relationships are particularly significant for this
discussion because they show up the real social power which the likes of
Guðmundr dýri and his peers had.
46
J. V. Sigurðsson, ‘Heimili, þingmennska, frændur og vernd á þjóðveldisöld’, in Íslenska
söguþingið 28.–31. maí 1997, I, 107–16 (p.111); S. Jakobsson, ‘Griðamál á ófriðaröld’, in ibid., I,
117–34 (p.127).
Page 100
Ordernr. 061297 P1
The connection of Ögmundr Þorvarðsson with Fnjóskadalur provides a very

interesting example of how some wealthy men could impose their power on some
valleys which were not very close to their home. Ögmundr’s reappearance in
Iceland after his extended absence in Norway was characterized by his ability to
use his sister’s farm (Draflastaðir) as a base for unreasonable actions against the
people of Fnjóskadalur. Although Guðmundar saga dýra seems to take against
Ögmundr, the kinds of crime he commits and the place he commits them may
suggest something of the geo-political realities. Ögmundr is alleged to have slept
with two women from families which actively supported Ögmundr’s father,
Þorvarðr, who lived in Fornastaðir (probably) and Laufás. In neither case is this
the kind of hospitality which the kinsmen of these women had in mind, even if
they were Þorvarðr’s supporters. The key thing is that these episodes show that
Þorvarðr had support in Fnjóskadalur, even though it is not obvious that he had
supporters in the districts between Hörgárdalur and Fnjóskadalur. Political
communities and political allegiances could span the landscape in ways which
might not be expected as a result of physical geography.
Fnjóskadalur is perhaps atypical, but there could clearly be clusters of support
for particular chieftains in particular districts which formed another layer of social
complexity. Elsewhere, there are more mundane examples of this. Guðmundr dýri
owned farms in upper Skagafjörður at one stage which no doubt owed allegiance
to him. In the case of both Fnjóskadalur and Guðmundr’s Skagafjörður farms it
is probable that political needs and economic needs combined. The earliest mál-
dagi for the church at Glæsibær,47 at the mouth of Hörgárdalur, owned woodland
in Fnjóskadalur which was a valuable resource unlikely to have been plentiful in
Hörgárdalur. Guðmundr’s Skagafjörður properties (Hálfdanartunga and Uppsa-
lir, ch.10) were no doubt better than many he could have had in Hörgárdalur.
As well as chieftains’ economic strategies which allowed a geographically more
dispersed political community to operate, there were more straightforward politi-
cal allegiances which clearly mattered in court and in the enforcement of legal
action with physical force. Kolbeinn Tumason’s regular support of Guðmundr
dýri was invaluable to the latter on several occasions; Sæmundr Jónsson (from the
south of Iceland), the Sturlungar clan (from the west), and others, served to
influence events at the Alþing (e.g. ch. 15) which had an impact at a regional level.
Thus the social networks of relatively small-time chieftains reached across Iceland,
even if they rarely stepped outside their own region.
47
Diplomatarium Islandicum, III, 520.
Page 101
Ordernr. 061297 P1
86 Chris Callow
Conclusions
The saga demonstrates a willingness of the most powerful figures in the district
to be at peace with most (less powerful) people because they obviously know each
other. Guðmundr dýri, about whom we know most, was moderate in his treat-
ment of some local people caught up in the Önundarbrenna. At the same time he
was keen to punish individuals whom he regarded as real enemies, such as Björn
Steinmóðarson. Kolbeinn Tumason, from outside Hörgárdalur, was more moder-
ate in his attitude to Björn but at the same time prepared to raid the valley on his
way out, not having to concern himself with the opposition of people who lived
close to him. The different reactions of Guðmundr and Kolbeinn serve to show
the varying kinds of social relations which appear in Guðmundar saga dýra:
Guðmundr had to be more cautious because he lived in a reasonably close-knit
community which could be reasonably well defined geographically.
Community in late twelfth-century Iceland had several dimensions. It was
indeed shaped by geography and distance from one’s home: the parish and/or
hreppur produced one form of social connection for everyone. At the same time,
the mobility of people at all levels of society cannot be underestimated — travel
and movement between households every year meant that individuals formed
relationships with a wide range of people, often outside their parish and outside
their blood relationships. The need for some kind of patron (goðorðsmaðr) for
most heads of households meant that they formed part of a wider political com-
munity too, which is what most saga action focuses on. People living tens, or even
hundreds, of kilometres apart formed part of the same interest group because, at
certain times, they shared a common enemy. While the most important kind of
community was the local one, formed by human and physical geography, the
reality was that life was far more complicated, and the relief map can never fully
explain the varied kinds of social connection which medieval Icelanders of all
levels of society actually had.
Page 102
Ordernr. 061297 P1
C OMMUNITIES OF D ISPERSED S ETTLEMENTS:

S OCIAL O RGANIZATION AT THE G ROUND L EVEL
IN T ENTH - TO T HIRTEENTH -C ENTURY ICELAND
Orri Vésteinsson
I
once worked as a research assistant in a project on the Icelandic family. The
project dealt with the Icelandic family in modern society,1 but my brief was
to look at the historical dimension, to see if some of the characteristics of the
modern phenomenon could be explained or put into context by its history.
One of the things I found — and one that has stuck in my mind since — is the
almost complete lack of courting rituals in early modern Icelandic society. This
is in stark contrast to places even like Norway where people also lived in separate
farms rather than villages.2 Indeed there was in early modern Iceland a distinct
absence of traditional or regular social gatherings where individuals could meet
people other than those of their own household or farm. Sunday Mass was of
course the major exception, but apart from that it seems that people, women in
particular, had very few occasions to travel beyond their own farm and the
majority of Icelanders would perhaps spend most of their lives in the company of
no more than five to fifteen people. While modern courting and dating rituals in
the West are basically urban and bourgeois customs they do ultimately derive
from the conditions of rural life, the villages and towns that created the frame-
work for the life of the majority of Europeans until the nineteenth century. In
contrast, modern urban and bourgeois Icelanders do not, as a rule, practise such
rituals and tend to just get on with it when it comes to mating. Whether one
1
S. Júlíusdóttir, Den kapabla familjen i det isländska samhället: en studie om lojalitet,
äktenskapsdynamik och psykosocial anpassning (Göteborg, 1993).
2
D. Gaunt, Femiljeliv i Norden (Stockholm, 1983), pp.23–40.
Page 103
Ordernr. 061297 P1
88 Orri Vésteinsson
considers this barbaric or pragmatic it seemed to me that the basic difference in

this respect between Iceland and more or less everywhere else was to do with
settlement patterns. That because Icelanders lived in isolated farms — with an
average household size of 6.15 in 17033 — and had very few occasions to socialize
with people from other farms, the conditions were simply never created for any
sort of courting rituals, especially not of the sort described for the Norwegian
countryside where large groups of young people paid social calls on each other.
There are a number of problems with this reasoning. Aristocratic courting, for
instance, is not dependent on geographic proximity; yet Icelandic aristocrats seem
not to have bothered any more than the general populace. So maybe the reason is
somehow more cultural. It is, for example, no doubt relevant that in the seven-
teenth century Church and state connived in successfully banning horse-fighting
and dances, both major occasions for gatherings of large groups of people, the
debauchery attendant on the latter, in particular, worrying the puritanical
authorities no end. So maybe it is the puritan influence on Icelandic culture that
is to blame.
I am not going to try to solve this problem here — suffice it to say that I think
that dispersed settlement can at most have been a contributing reason to the
absence of courting rituals in modern Icelandic culture. Indeed I want to use this
occasion to argue the opposite: that in the Viking Age and high medieval times at
least, Icelanders lived in well-defined communities and that while the settlement
patterns of these communities were often quite dispersed they had a cohesion that
betrays their conceptual origin as hamlets or villages.
The Archaeology of Communities
One of the problems facing archaeologists who excavate settlements is defining

the social context of their settlements. In the cases of villages or towns it is nor-
mally taken for granted that each of them constituted a community — a place
where three or more families live in close proximity must in fact constitute a com-
munity in some sense. It may not, however, be the whole story. Small settlements,
hamlets or small villages, may for instance belong to much larger communities,
which may be — in a number of ways — more significant for their inhabitants
than their day-to-day dealings with their closest neighbours. Likewise large villages
3
Hagskinna: Sögulegar hagtölur um Ísland, ed. G. Jónsson and M. S. Magnússon (Reykjavík,
1997), p.140.
Page 104
Ordernr. 061297 P1
COMMUNITIES OF DISPERSED SETTLEMENTS 89
and towns may consist of multiple communities, where individuals and families
living within earshot of each other may have little or no common interests or
dealings. In both these cases, hamlets on the one hand and towns on the other, the
obvious archaeological definition of community, based on the proximity of dwell-
ings to each other, may not at all be the most significant for understanding the
working of the settlements.
In the cases of isolated, single household, settlements, archaeologists are nor-
mally more aware of the lack of social context, although it is not unfair to say that
this does not as a rule seem to cause them much concern. There is, however, every
reason to be concerned about this as agricultural settlements cannot exist in
complete isolation. It is arguable that they must always belong to some larger
community and that without an understanding of this broader context the
archaeologists will not be able to appreciate properly the nature of the settlement
they have unearthed.
For medieval Europe documents provide the general background and archae-
ologists can assume that their settlement belonged to a parish* and that it had
some tenurial, fiscal or other ties of obligation with places elsewhere. Medieval
archaeologists sometimes try to relate individual excavated settlements to a parish
church and seigneurial centre, but rarely are archaeological criteria used to try and
define the limits of communities — to define which other settlements belonged
to the same community as the one under study. In contrast, this sort of work is
commonplace in many areas of prehistoric archaeology, where emphasis is placed
on identifying settlement hierarchies and describing the extent and possible
interrelationships of the settlement units.4
Without place-name evidence or written sources prehistoric archaeologists are,
however, faced with severely limited opportunities to understand the nature of
the interrelationships between settlements of different ranks or tiers. They tend
to rely on fairly simplistic models developed by historical geography to rank
settlements. Central place theory and other such tools are used to create a sense
of a political landscape where large settlements/public architecture/rich artefact
collections are seen to equal centres of power and authority, and the relative
absence of these features are seen as indicators of dependence and subordination.
The number of definable tiers of settlement types is then seen as an indicator of
social complexity. Conventional archaeological methods cannot do much more
4
See e.g. discussion in J. Yeager and M. A. Canuto, ‘Introducing an Archaeology of
Communities’, in The Archaeology of Communities: A New World Perspective, ed. M. A. Canuto
and J. Yeager (London and New York, 2000), pp.1–15 (pp.3–5).
Page 105
Ordernr. 061297 P1
than that, yet it is obvious that they can only produce the most general picture of
the interrelationship of settlements, and will indeed not pick up on most of the
significant patterns in the settlement structure of any given society. Historical
archaeology is much better placed to study such patterns, and should be doing so,
not only to understand better the historical societies it deals with, but also because
a systematic study of the archaeological, historical and toponymic evidence for
settlements should result in a better general understanding of settlement patterns
which could then be applied to the study of prehistoric societies. Without offering
any firm results or simple solutions, the following is intended as a contribution to
such an effort.
North Atlantic Settlement Patterns
It is commonplace in Scandinavian archaeology and settlement studies to point

to the fact that settlement patterns change fundamentally along the border where
coniferous forest takes over from the deciduous woodlands of southern Scandi-
navia. In Denmark, southern Sweden and the very southernmost part of Norway,
settlement patterns are dominated by villages; whereas in the northern — conif-
erous — areas settlement patterns are more dispersed, characterized by single
farmsteads.5 The latter pattern is also found in the largely treeless North Atlantic
islands colonized by Scandinavians in the Viking Age: the Northern Isles of Scot-
land, the Faroes, Iceland and Greenland. In the most general terms it is obvious
that this pattern is dictated by the environment: the less productive lands of the
north cannot sustain the same density of population as in more southerly lati-
tudes, resulting in smaller and more dispersed settlement units. This generaliza-
tion is, however, only useful in the broadest discussion of settlement types and
social structure. If one wants to take a closer look at individual regions and
attempt to understand the actual relationship between settlement patterns and
social structure in particular localities, it becomes an obstacle to understanding.
A good example of how fundamental the idea about isolated settlements in the
north is, and at the same time how wrong it can be, is provided by Faroese
5
Needless to say this generalization hides a complex and highly debated issue, well represented
by a collection of papers under the collective heading ‘Together or Apart – The Problem of
Nucleation and Dispersal of Settlement’, in C. Fabech and J. Ringtved, eds, Settlement and
Landscape: Proceedings of a Conference in Århus, Denmark May 4–7 1998 (Moesgård, 1999),
pp.125–202.
Page 106
Ordernr. 061297 P1
settlement patterns.6 The Faroes are a small archipelago midway between Shet-
land and Iceland, comprising some seventeen inhabited islands, most of which rise
sharply from the sea with only very limited lowland along the coasts, mostly in
short coastal valleys. The Faroes were settled by the Norse in the ninth century,
and by the sixteenth century the colony consisted of some eighty-five settlements
(pl. bygðir*), each comprising one or more fyrndarbýlingur* (sing. ‘old farm’)
which in turn could be subdivided into farmsteads and households. Many of the
bygðir had their own church and constituted separate parishes but many did not
— at least not in early modern times. Those that did not belonged to the parish
of a church in a neighbouring, often quite distant, bygð (sing.). The concept of the
fyrndarbýlingur is obviously related to the Icelandic one of lögbýli* — a taxable
unit of a certain value which can be bought and sold and may contain any number
of farmsteads and households, although the ideal was clearly that there was only
one main household per lögbýli. The fyrndarbýlingar (pl.) were the basic concep-
tual units in the Faroese settlement system, at least from the point of view of
landowners and the administration. For the archaeologist or geographer looking
at the map it seems more natural, however, to consider the bygð as the most
important unit of settlement. A bygð is in most cases a sharply defined stretch of
lowland typically occupied by one to four fyrndarbýlingar and a much greater
number of households, at least in early modern times. The farmsteads in a bygð
are, as a rule, not ordered in any particular way but scattered around an infield
defined by a common boundary — a head dyke. Each bygð typically consisted of
scores, if not hundreds, of people and can easily be classified as a hamlet or small
village. The bygð had defined functions, especially in regards to the management
of the outfield which was shared proportionately between the farms in each bygð.
Yet it was the fyrndarbýlingur which became the basic unit for taxation and prop-
erty transactions, a unit which seems almost artificial as it reflects neither the
geographical unity of the bygð nor the actuality of the farmsteads and households.
Each fyrndarbýlingur represents a grouping of farmsteads and households, nor-
mally into one main — conceived as original or central — farmstead and a varying
number of satellite households of lower social and economic status. The idea is
obviously that each bygð originally only had one, two or three fyrndarbýlingar
6
A. Thorsteinsson, ‘On the Development of Faroese Settlements’, in Proceedings of the Eighth
Viking Congress Århus 24–31 August 1977, ed. H. Bekker-Nielsen, P. Foote and O. Olsen
(Odense, 1981), pp.189–202; S. V. Arge, ‘Føroysk búsetingarsøga – tær fornfrøðiligu heimildir-
nar’, in Frændafundur, vol. II, ed. T. Sigurðardóttir and M. Snædal (Tórshavn, 1997), pp.41–58;
S. V. Arge, ‘Í Uppistovubeitinum: Site and Settlement’, Fróðskaparrit, 45 (1997), 27–44.
Page 107
Ordernr. 061297 P1
which then became subdivided at some undisclosed date after the settlement
period and before early modern times.
This notion of original large farm units which later have become subdivided
is of very ancient standing in the Nordic world, and has influenced both medieval
administrative practices as well as modern scholarship — not least in Norway
where place-name studies have attached great significance to this idea.7
It is of course significant that in the Faroes — as well as in Norway and Iceland
— the basic proprietary and fiscal units do not necessarily coincide with the
physical conglomerations of farms. That, however, is no different from anywhere
else and it only represents one aspect of the settlement system, and for the pur-
poses of people’s everyday lives and relations probably one of the less significant.
The Faroese settlement pattern is fairly homogeneous, much more so than the
varied, and more generally dispersed, settlement structure of Iceland. It serves
well, though, as an introduction to the Icelandic patterns as it shares all the basic
features of the latter.
Icelandic Settlement Structure
In order to define the settlement structure in Iceland two issues need to be dealt
with. First a definition of the physical relationships between settlements has to be
established. Is the settlement pattern really dispersed? How dispersed is it? How
much variation is there in the dispersal? What is an isolated farm and how isolated
is it really? Second, indications for groupings and associations of settlements need
to be recognized and characterized.
To start to deal with the first issue, eight study areas were selected based on a
rough division of Iceland’s inhabited areas into four settlement types: valley, plain,
fjord and coast.8 Two 12x12 km areas were selected to represent the range within
each type. The results are summarized in Table 5.1 and the patterns are shown in
Fig. 5.1. This sample is neither comprehensive nor random, but it does at least
provide a sizeable dataset that can serve as the basis for the argument of this essay
— more work is obviously needed for a full treatment of Icelandic settlement
patterns. In this discussion use is made of early modern tax-value and census data
to qualify the geographical information. This is based on the widely accepted
7
E.g. M. Olsen, Farms and Fanes of Ancient Norway (Oslo, 1928).
8
O. Vésteinsson, The Christianization of Iceland: Priests, Power, and Social Change
1000–1300 (Oxford, 2000), pp.12–15, describes the first three types.
Page 108
Page 109
Table 5.1. Geographical and demographic data for the eight study areas shown on Figure 5.1.
Ordernr. 061297 P1
Total area habitable km2

No. of households in 1703
No. of farms in 1695

No. of people in 1703
Mean household size in 1703
km2 per farm in 1695
km2 per household in 1703
No of people per km2 in 1703
neighbour in m
Closest distance to nearest
neighbour in m
Greatest distance to nearest
neighbour in m
Mean distance to nearest
Area
Hornstrandir – dispersed fjord settlement 22.3 5 7 43 6.14 4.46 3.18 1.93 2516 4091 3254
Dýrafjörður – dense fjord settlement 55.2 23 60 338 5.63 2.40 0.92 6.12 162 2794 986
Öxnadalur – planned valley settlement 17.4 11 10 60 6.00 1.58 1.74 3.45 551 2258 943
Eyjafjörður – dense valley settlement 57 40 57 320 5.61 1.43 1.00 5.51 172 1931 710
COMMUNITIES OF DISPERSED SETTLEMENTS
Úthérað – dispersed plain settlement 121 17 23 172 7.48 7.11 5.26 1.42 1153 3816 2039
Ölfus – dense plain settlement 99.9 35 89 558 6.27 2.85 0.89 5.58 282 2526 732
Grindavík – nucleated coastal settlement (62.3) 7 34 185 5.44 (8.90) (1.83) (2.96) 720 3513 1449
Innnes – dense coastal settlement 78 64 192 1053 5.48 1.22 2.46 13.5 157 2281 628
Totals/Averages 513.1 202 472 2729 5.78 2.54 1.09 5.32
Notes
Habitable area is defined as all dry land (i.e. excluding sea, rivers and lakes) under 200 m above sea level. In the case of Grindavík the
largest part of this area is lava fields with limited vegetation. That means that the areal figures for that area are not comparable to the
others. The number of farms is based on a register from 1695 (edited in Björn Lárusson, The Old-Icelandic Land Registers) which lists
all the lögbýli in Iceland, preferred over the 1703 census which is household-based and is not concerned in the same degree with
defining the status of farms. Household numbers and population figures are based on the census. Nearest-neighbour measurements
are taken between farms of the 1695 register from the 1:100,000 scale maps of Iceland published by Landmælingar Íslands. The
93
accuracy of the measurements is within 50 m.

Ordernr. 061297 P1
Page 110
Ordernr. 061297 P1
Figure 5.1 (left and above). Eight study areas representing different settlement
types in Iceland: fjord (Hornstrandir, Dýrafjörður), valley (Öxnadalur,
Eyjafjörður), plain (Úthérað, Ölfus) and coast (Grindavík, Innnes).
assumption that settlement patterns remained largely unchanged from the settle-
ment period in the ninth to tenth centuries until the onset of industrialization in
the late nineteenth century.9 If this is true, the early modern data can be used to
give approximate ideas about conditions in the Viking Age and High Middle
Ages. Although reservations must be made for individual cases, this assumption
9
B. Lárusson, The Old-Icelandic Land Registers (Lund, 1967), pp.32–33; O. Vésteinsson,
T. H. McGovern and C. Keller, ‘Enduring Impacts: Social and Environmental Aspects of Viking
Age Settlement in Iceland and Greenland’, Archaeologia Islandica, 2 (2002), 98–136
(pp.117–18).
Page 111
Ordernr. 061297 P1
is supported here in order to bring out the main characteristics of Icelandic

settlement patterns.
While this is not apparent from the study areas (although see discussion of
Grindavík below) there were villages — in the sense of a clustered settlement with
multiple farms and households — in Iceland before the nineteenth century. There
were not many, though, and they were in no way typical of Icelandic rural life.
They are important in the context of this discussion, however, as they show which
conditions were needed for dense settlement to develop in Iceland. Icelandic pre-
industrial villages are of two types. One is the fishing village found in a few off-
shore islands, like Vestmannaeyjar (330 souls in 56 households in 170310), Flatey in
Breiðafjörður (106 in 21 households), Bjarneyjar in Breiðafjörður (85 in 13
households) and in Grímsey (87 in 17 households), and in coastal settlements like
Stokkseyri (140 in 27 households), Selvogur (114 in 23 households), Arnarstapi (147
in 32 households), Hellnar (194 in 42 households), Rif (106 in 26 households) and
Hellissandur (243 in 61 households). Except for Grímsey all these major con-
centrations of people are found in the south-west and west (Fig. 5.2). Many of
these villages were also seasonal fishing stations, the population of which might
multiply several times in late winter and early spring.11 Of the other type of pre-
industrial village there is only the one example, the farming village Þykkvibær in
Holt (68 persons in 15 households; for the meaning of bær*, see glossary). Þykkvi-
bær is distinguished from other rural clusters of households (see below) not by its
size but by having a village lay-out (a double line of farms on either side of a street,
although probably organic rather than planned) and no central household, in fact
not even a church. For the fishing villages at least it is normally postulated that
they are an early modern phenomenon, the consequence of the rise of commercial
fishing in the thirteenth and fourteenth centuries. This is by no means certain
however — commercial fishing may have started much earlier and fishing does not
need to be commercialized to become the basis for subsistence of large groups of
people. About the population of Þykkvibær in the Middle Ages nothing is known,
although the most plausible explanation for the place-name itself is that it refers
to the density of the settlement (from adj. þykkur ‘thick, dense’)12 which can be
taken as an indication for the antiquity of this village.
10
Manntal á Íslandi árið 1703 tekið að tilhlutan Árna Magnússonar og Páls Vídalíns
(Reykjavík, 1924–47).
11
L. Kristjánsson, Íslenzkir sjávarhættir (Reykjavík, 1982), II, 34–63.
12
Á. Óla, Þúsund ára sveitaþorp: Úr sögu Þykkvabæjar í Rangárþingi (Reykjavík, 1962), p.30.
Page 112
Ordernr. 061297 P1
REYKJAHVERFI
MÝVATNSSVEIT
ES
SN EI
-D
ÍM SK
GR
Melar = farms
mentioned in the text
Figure 5.2. Iceland. Places and regions mentioned in the text.
Such villages are, however, clearly anomalous and most Icelanders in the
Viking Age and medieval times lived in more dispersed settlements. The eight
study areas can be divided into two, based on the distance of a farm to its nearest
neighbour. These vary enormously, from 157 m to more than 4 km, but in the
majority of the study areas the average distance between neighbouring farms is
within 1 km. The three study areas which have longer average distances than 1 km
represent only 15% of all the farms and can all be considered anomalous, albeit for
different reasons.
Hornstrandir represents an area of truly isolated farms, where not only are dis-
tances between settlements great but the roads are also very difficult and most
transport was by sea. Many of the farms in this and similar areas were not perma-
nently occupied and such areas were the first to be abandoned in periods of
economic downturn. Of the five farms occupied in this area in 1703 one had only
two souls and one had four, whereas in another (Furufjörður) there were three
households with twenty-three persons between them. This type of settlement is
only found in the north-west (Vestfirðir), far north (e.g. Héðinsfjörður, Fjörður,
Náttfaravíkur), north-east (Langanes) and eastern seaboard (e.g. Húsavík, Breiða-
vík, Sandvík), and represents at most 3% of all farms in the country. Grindavík is
Page 113
Ordernr. 061297 P1
Photo 5.1. Fjord settlement. Dalatangi, E-Iceland. Photo: Oddur Sigurðsson.
also an anomalous type of settlement, but of a very different kind. Here the house-
hold and population figures show that each farm is really a hamlet of its own:
Járngerðarstaðir with forty-four persons in eleven households; Þórkötlustaðir
with thirty persons in six households; and Staður with forty persons in five house-
holds. Each of these clusters was commonly called hverfi* (e.g. Staðarhverfi), one
of the Icelandic terms for neighbourhood or hamlet.
Unlike Innnes, where post-medieval fragmentation accounts in part for the
large number of farms in 1695, in Grindavík the farm units have been unchanged
since the fourteenth century at least, but it is not known whether the farms were
as densely populated in medieval times as they were in the beginning of the
eighteenth century. For Grindavík the nearest neighbour figures therefore do not
give a balanced idea of social relations. Quite unlike the people of Hornstrandir,
the people of Grindavík lived in densely populated clusters separated by flat — if
barren — land. In terms of farms the Grindavík settlement pattern is even rarer
than the Hornstrandir one, representing less than 1% of all the farms in the
country. In terms of population, however, the proportion is presumably con-
siderably higher.
Page 114
Ordernr. 061297 P1
Úthérað is the third study area with long distances between farms and the only
one of the three that can be considered as representative for a significant number
of Icelandic settlements, possibly some 10% to 20% of all farms in the country.
Similar patterns are found in the leaner parts of many plains environments in
Iceland: for example, Mýrar in Borgarfjörður; Grímsnes and parts of Skeið and
Flói in Árnesþing (for þing*, see glossary); parts of Holt, Landssveit and Rangár-
vellir in Rangárþing; Ásar in Húnaþing; the flat coastal areas of the north-east
such as in Bakkafjörður and Þistilfjörður; and highland plateau settlements like
Hólsfjöll and Mývatnssveit. Many of these areas are also characterized by low
taxation values, presumably reflecting the limited productivity of such flatlands.
They are also characterized by single household farms, with five to ten souls on
each farm.
In settlements of this type communications were as a rule relatively easy
although the distances are often quite long. The main obstacles tend to be large
rivers and extensive bogs, barriers that pose greater problems in summer than
winter. Unlike Hornstrandir where the farms could be effectively isolated for
months during the winters, in areas like Úthérað communications will have been
easiest once the bogs and rivers froze over, making winter, rather than summer,
the time when people could meet more easily. This may have been quite signifi-
cant as winter effectively represents more than half the year and is also the time
when people had the greatest leisure. Úthérað therefore represents a truly dis-
persed settlement with small single family farms separated by considerable dis-
tances. There is some clustering — one in the lower left-hand and the other in the
upper right-hand corner of the map — but even in these the distances between the
farms are always more than 1 km.
In the five other study areas the settlement is denser and in places so dense that
it is justifiable to classify the clusters as hamlets. It is, however, important to note
that even in the areas of densest settlement, like Innnes, Ölfus and Eyjafjörður,
there are a few farms that are located 2 km or more from their nearest neighbour.
In these areas of denser settlement two main types of patterns can be discerned.
Öxnadalur represents a good example of a planned settlement with a string of
farms evenly distributed on either side of a narrow valley.13 As in Úthérað these
are all single household farms with low taxation values, in fact more uniformly so.
In the other areas similar, but shorter, strings of more or less uniformly small
13
O. Vésteinsson, ‘Patterns of Settlement in Iceland: A Study in Prehistory’, Saga-Book of the
Viking Society, 25 (1998), 1–29 (pp.20–23); Vésteinsson, McGovern and Keller, ‘Enduring
Impacts’, p.121.
Page 115
Ordernr. 061297 P1
single household farms are also found. In Innnes they are found in the lower left-
hand quadrant of the map whereas in the remaining three areas, Dýrafjörður,
Eyjafjörður and Ölfus, they cannot be discerned from the spatial arrangement
alone. In Eyjafjörður the majority of the farms have single households and there
is hardly any clustering, but differentiation is evident in the taxation values and
variations in the size of the holdings. Many of the farms with low taxation values
have limited access to mountain pasture and therefore probably represent planned
settlements.14
In Dýrafjörður, Ölfus and Innnes the patterns are characterized by clustering.
In Dýrafjörður most of the farms have multiple households, and two have as many
as seven, with more than thirty persons on each. In the lower left-hand corner of
the map there are two clusters made up of three farms each. These are all small
farms in terms of taxation values but they all have multiple households, so that
one cluster has seven households and thirty-nine souls whereas the other has five
households and twenty-six souls. These two clusters may represent satellites from
Sandar, a major estate15 with a parish church (sóknarkirkja*, cf. sókn*) at the con-
fluence of the two rivers draining the valleys in which the two clusters are found.
These clusters are reminiscent of the Faroese settlement patterns discussed
above, where settlements are closely confined by topography but where farming
units have been defined on some completely different premises. In Ölfus such
clusters of small farms with multiple households are also found, especially towards
the upper left-hand corner of the square. A different type of cluster is found on
either side of the large river running across the lower half of the square. On the
northern bank there is a cluster of four farms which is in fact a village of sixteen
households and ninety-four souls. At the core of this village there is a major estate,
Arnarbæli, to which most of the households belong. On the periphery there are
two much smaller farms, one single household farm and one with three house-
holds and twenty-three persons. The last mentioned, Auðsholt, had its own
chapel* (bænhús*, referred to in the fourteenth century16 but presumably dating
back to the eleventh) and therefore presumably constitutes a unit the indepen-
dence of which vis-à-vis Arnarbæli must be of considerable antiquity.
14
Vésteinsson, ‘Patterns of Settlement’, p.22, Fig. 8, shows the subdivisions on the western
side of the valley.
15
‘Estate’ is here used to describe large holdings with multiple households, of which one is
central, usually associated with a church.
16
Reykjavík, 1853–1976), IV , 92.
Page 116
Ordernr. 061297 P1
Photo 5.2. Valley settlement. Fnjóskadalur, N-Iceland. Photo: Oddur Sigurðsson.
On the south bank of the river a cluster of three farms is the more centralized
estate of Kaldaðarnes with thirteen households and seventy-seven persons. The
estate also included the single farm at the lower edge of the map which is, in fact,
a small cluster of five households and twenty-five persons. Similar clusters are also
found in Innnes, but in Álftanes, the more southerly of the two main peninsulas
shown on the map, there is also a very tight cluster of a slightly different kind. In
the cluster of eight farms on the northern side of the inlet, which nearly cuts the
peninsula off from the mainland, no one farm or household can be considered as
the core and all can be described as modest holdings. Together they had twenty-
four households and 138 persons in 1703.
To sum up the discussion so far we have seen that Icelandic settlement pat-
terns are far from homogeneous. There is considerable range within the country,
from dense settlements that can only be described as villages to truly isolated
farms. Most Icelanders lived, however, in neither type of settlement but in farms
of one to two households within 1 km of the nearest neighbour. One kilometre
hardly represents isolation — a ten- to fifteen-minute walk in normal terrain —
and is no hindrance to people who are intent on meeting each other. It is,
Page 117
Ordernr. 061297 P1
however, a significant distance in that it restricts accidental meetings and limits

the degree to which neighbours can monitor what is going on in other households.
The social life in such settlements is therefore significantly different from that of
villages. Whether settlements of this type represent communities is the issue we
turn to next.
Reconstructing Icelandic Communities
There are several different types of evidence that can be brought to bear on the
problem of community structure in Iceland. Those that will be considered here
are place-names, parish and commune organization, and property associations. Of
these the place-names presumably represent the earliest layer of evidence. Most
Icelandic farms seem to have been in continuous occupation since the settlement
period and to have retained their names throughout — the exceptions to this rule
have been identified and described by Ólafur Lárusson.17 The parish and to a
lesser extent commune (see hreppur*) boundaries can in many places be recon-
structed for the twelfth to thirteenth centuries and can be considered to also
reflect the late or immediately post–Viking Age community structure. Informa-
tion on ownership is more problematic as comprehensive data only exists from
early modern times, but it will be argued here that the structure of church
properties, charter evidence for which is available from the thirteenth and four-
teenth centuries, can be considered as representative of late Viking Age settlement
patterns.
Icelandic has a few concepts for settlement clustering, ranging from þorp*
(‘village’), and the much more commonly used hverfi (‘hamlet, neighbourhood’),
to the more general byggð* (‘settlement’) and sveit (originally ‘group, company of
men’; in a spatial context, ‘rural area’, ‘countryside’). Þorp does occur as a descrip-
tion of clusters of farms, particularly in the north-west (e.g. Vaðalsþorp), but there
are few, if any, medieval instances of such a usage for Icelandic settlements. Hverfi
on the other hand is attested in medieval records both as a description of large
areas (e.g. Kelduhverfi, Reykjahverfi) and of individual clusters of farms. An exam-
ple of the latter is Fiskilækjarhverfi mentioned in two thirteenth-century sagas18
17
Ó. Lárusson, ‘Úr byggðasögu Íslands’, in Byggð og saga (Reykjavík, 1944), pp.9–58
(pp.41–50).
18
Eyfirðinga so3 gur, ed. J. Kristjánsson, Íslenzk fornrit, 9 (Reykjavík, 1956), pp.54–55;
Ljósvetninga saga, ed. B. Sigfússon, Íslenzk fornrit, 10 (Reykjavík, 1940), p.235.
Page 118
Ordernr. 061297 P1
Figure 5.3. Place-name evidence for neighbourhoods in Iceland.
and shown on the map of Eyjafjörður (Fig. 5.1). It is not explained in the saga
exactly which farms belong to this unit, but from considering the map and the
place-names it is apparent that this name can only have applied to a group of four
or five farms on either side of the quite humble brook Fiskilækur — a geographical
feature of such insignificance that it can hardly have given a much larger area its
name. These four or five farms do not represent a particular clustering, but the
name given to them as a group suggests that they must have constituted a com-
munity — a hamlet — with a separate identity from the neighbouring farms.
In an area now called Garður in the south-west there is a major farm called
Útskálar (Fig. 5.3). Some 3.7 km to the south-east there is another smaller farm
called Gufuskálar. The area in between is quite densely settled and in a late
thirteenth-century document this is referred to as Miðskálagarður.19 The com-
mon element to all these names is the term skáli (‘building, hall’) and út and mið
are directional concepts meaning ‘out (more westerly, closer to the sea)’ and
19
Diplomatarium Islandicum, II, 78–80.
Page 119
Ordernr. 061297 P1
‘middle’ respectively. It seems then that this neighbourhood of some forty house-
holds and 182 persons in 1703 was originally called *Skálar20 implying a single
community. This seems to have been conceived of as having three sub-clusters:
one major estate with parish church and many dependent households (Útskálar);
one cluster of modest holdings (*Miðskálar); and a single small estate somewhat
separated from the other two (Gufuskálar). The first two retained their com-
munal identity in the presumably more recent term Garður, whereas Gufuskálar
was in later centuries considered to belong to the area of Leira along with other
farms further south.
In Ólafsfjörður in north Iceland there are two adjacent farms now called
Ósbrekka and Skeggjabrekka, separated by some 620 m (Fig. 5.3). According to
Guðmundar saga dýra these farms were both occupied in the late twelfth century;
they had adjoining homefields and were both called Brekka.21 As the purpose of
place-names is surely to identify significant elements in the landscape it makes
little sense to call adjacent farms the same name unless they are considered to be
parts of the same unit. Brekka means ‘slope, hillside’ and is an apt description of
the landscape in which the farms are situated. They are located on the western side
of a large lake on the opposite side of which there is also a hillside with two further
farms called *Brekka (now Bustarbrekka and Hornbrekka). All these farms share
a common location on sloping, in fact quite steeply sloping, ground, in contrast
to other farms further inland that are all located on the valley floor. It is surely this
contrast which has occasioned the name, but the fact that all four farms share the
same topographic qualifier could suggest that they were originally conceived of as
a unit, and therefore possibly a community.
In Aðaldalur in north-east Iceland there are three farms all called *Hvammur
(Fig. 5.3). The northernmost is Ystihvammur (‘outermost-’ ), then there is Mið-
hvammur (‘middle-’ ) and furthest south is Presthvammur (‘priest’s-’), the last
name no doubt originating in the late Middle Ages,22 presumably because it was
owned or occupied by a priest at some point. Its original name is likely to have
been *Syðstihvammur (‘southernmost-’ ) or *Fremstihvammur (‘innermost-’). In
many cases where adjacent farms share the same name and are distinguished by
directional elements (in–out, up–down, north–south, etc.) or by dimensional
elements (large–small) it is because an original single farm has become split into
20
An asterisk in front of a place-name (as here and below) denotes a form that does not occur
in any source, but can be reconstructed.
21
Sturlunga Saga, ed. Ö. Thorsson (Reykjavík, 1988), pp.143–44.
22
Diplomatarium Islandicum, V , 111.
Page 120
Ordernr. 061297 P1
two or more, often in early modern times. In cases where adjacent farms are
attested under the same name in early records and/or where such farms are
separated by other farms with different names it seems, however, that such name-
sharing farms may be of similar antiquity. In this case, Presthvammur and
Miðhvammur are separated by two farms, Brekka and Klömbur, and the whole
area, including the farms Hraun and Hagi was (and still is) called Hvammar
(plural of Hvammur).23 In the thirteenth-century Reykdæla saga a man is intro-
duced as living in Hvammur, suggesting that the common identity of the cluster
was still more significant at that date than that of the individual holdings.24
Hvammur means ‘hollow’ and clearly describes the dominant landscape feature,
a broad and curving hollow into the hillside at the foot of which the farmsteads
stand. It seems that initially each of the farmsteads did not have a separate name
and that the whole area of fairly dispersed farmsteads was called *Hvammur,
suggesting that this fairly well-defined geographical area also constituted a com-
munity with an identity separate from neighbouring farms.
More concrete evidence comes from Neðri-Ás in Hjaltadalur, the central farm-
stead of a large estate no doubt originally called *Ás. Here the excavation of an
eleventh-century chapel and cemetery revealed that the population using the
cemetery is likely to have been considerably larger than the household of the
central farmstead. The cemetery was used by a group of thirty to thirty-five peo-
ple, and the layout of the graves is suggestive of family or household plots within
the cemetery. It is suggested that during the eleventh century this cemetery was
used by at least five households, all located within what later became the farms
Neðri-Ás and Efri-Ás, reflecting their close association and the existence of an *Ás
community.25
It is quite common that names of major estates and parish centres are derived
from geographical or other features that, rather than being particular to that
individual farmstead, describe a larger area (e.g. Melar in Melasveit, Mælifell in
Skagafjörður, Garðar in Álftanes). Although it is difficult to prove it seems that
such names were originally names for territories, and hence communities. In such
cases it seems that the political and economic centre of the community has
retained the name but that lesser holdings have acquired new names, often with
personal names rather than topographic names as their first element.
23
E.g. Diplomatarium Islandicum, V , 34.
24
25
O. Vésteinsson, Forn kirkja og grafreitur á Neðra-Ási í Hjaltadal (Reykjavík, 2000).
Page 121
Ordernr. 061297 P1
Parishes as Communities
In the cases of Fiskilækjarhverfi, *Brekka and *Hvammur discussed above, the

communities suggested are not clustered at all. They are quite dispersed and
without the place-name evidence would not be obviously distinguishable as com-
munities with separate identities. They are also quite small and may all have been
originally single holdings. Therefore, either they may represent very early commu-
nities, which lost their cohesion when the holdings were divided up and were
given individual names, or they can be seen as some sort of sub-communities,
neighbourhoods within more structured communities. It is to such structured
communities we now turn.
Parishes (sóknir) developed in Iceland in the course of the twelfth and thir-
teenth centuries.26 Each church had a tithe area* — sometimes only the farm it
was attached to — but those churches which did not have a resident priest were
considered to belong to the parish of one that did. In addition there were chapels
to which no tithes or other dues were paid. On this basis it is possible to propose
a ranking of farms according to their relationship to a church or chapel. The
major distinction is between those farms that had no chapel or church and those
that did. The other distinction is between farms with different types of churches.
The former distinction represents differences in rank between farms in the early
to mid-eleventh century, following the conversion around 1000, when it seems
that most churches and chapels were originally built. In this early period of church
building it does not seem that there were any significant differences in the sizes or
ranks of the churches — they were basically all very small, and presumably private,
chapels. Significantly, all the churches and chapels are associated with cemeteries,
suggesting that community burial was one of the principal functions of the
eleventh-century chapels/churches.27 Churches and chapels are as a rule found on
farms with high taxation values, so the fact that some farmers built churches on
their farms and some did not seems to reflect status: the occupants of farms
without a church or a chapel probably represent the dependent part of the popula-
tion, tenants and unfree farmers.
26
Vésteinsson, Christianization of Iceland, pp.101–07, 238–45; O. Vésteinsson, ‘Íslenska
sóknaskipulagið og samband heimila á miðöldum’, Íslenska söguþingið 28.–31. maí 1997:
Ráðstefnurit, 2 vols (Reykjavík, 1998), I, 147–66 (pp.153–59).
27
O. Vésteinsson, ‘The Formative Phase of the Icelandic Church c. 990–1240 AD ’, in Church
Centres, ed. H. Þorláksson (Reykjavík, 2005), pp.71–81.
Page 122
Ordernr. 061297 P1
The formation of the parish system, taking off in the twelfth century, further
accentuated the differences within the presumably free-holding farming popula-
tion by creating a ranking system. In this a relatively small number of churches
became parish centres and an equally large group retained certain rights and a
degree of independence vis-à-vis the parish church to which they were neverthe-
less subordinate. The majority of the chapels, however, became completely
dependent on services from the parish church to which they also had to pay dues.
The ranking evident in these differences allows us to distinguish four groups of
farms, and the way parishes formed around these is quite revealing of the nature
of rural communities in Iceland in the wake of the Viking Age.
Parishes can be divided into two main types. The first is the estate parish where
the parish is either made up of a single estate or has a large estate at its core and
smaller ones and individual farms on its periphery. Single estate parishes are the
exception and in the study areas only two are found, both in Innnes — those of
Nes and Reykjavík, with thirty-five and forty households in areas of approximately
3 and 10 km 2 respectively. Before these estates became royal property they were
owned and managed by a single landlord each, as far as can be seen members of the
aristocracy, who owned the church and hired the priest. The tenants and
parishioners therefore had much in common, and many occasions to meet and
share their troubles.28
Most estates were significantly smaller, typically with five to ten households
on the central farm and directly associated holdings (see below). In the case of
Laugarnes, also on the Innnes map, there were five households on the estate in
1703, and no directly associated holdings, while ten other farms with sixteen
households belonged to the parish.29 The furthest from Laugarnes — which is not
at all centrally located in the parish — was more than 10 km away.
Another similar example is Reykir in Ölfus. Here there was only one house-
hold at the central farm itself but six more on the estate, which was quite extensive
with more than 2.5 km from the cottage furthest out to the central farm. On
either side of the estate there were ten or eleven farms with sixteen households in
1703 that belonged to its parish and paid tithe to Reykir.30 One of these farms,
Núpar, had its own church, the owner of which retained his household’s tithe pay-
28
Ó. Lárusson, ‘Hversu Seltjarnarnes byggðist’, in Byggð og saga, pp.84–122 (pp.114–18);
Vésteinsson, ‘Íslenska sóknaskipulagið’, pp.162–64.
29
Diplomatarium Islandicum, XV , 636.
30
Diplomatarium Islandicum, XII, 662; AM 66a 8 vo , pp.42a–43a.
Page 123
Ordernr. 061297 P1
ments.31 This farm is set apart from all the others, in terms of both location and
taxation value, and seems to represent a small estate which would have been the
next in line to become a parish centre after Reykir and Hjalli on either side of it.
The second type of parish is where the parish church is located on a relatively
humble farm, not the area’s major estate (which is quite rare), or in areas where
there does not seem to have been any single major estate. An example of the latter
is Hjaltastaðir in Úthérað. Hjaltastaðir was a larger than average farm but did not
compare in the quality of its land to the three farms to its north (off the map),
which represent a cluster of equally large holdings occupying the best land avail-
able in the area.32 One of those farms had its own church that retained its own
tithes but was serviced from Hjaltastaðir, and there were a further six chapels in
the parish which had a total of nineteen farms.
The parish of Hjaltastaðir is an area of very dispersed settlement, the structure
of which nevertheless suggests that it can be thought of as a community, whose
inhabitants considered each other to be neighbours with a range of common
interests. The establishment of the parish church at Hjaltastaðir seems to be the
result of a communal effort, witnessed most clearly by its somewhat marginal loca-
tion and the fact that all the farms in the parish paid an extraordinary due to the
church, the feeding of a lamb annually.33 Such dues can only have been established
in the first place if the whole parish was owned by the same individual or family
— which seems unlikely although not at all impossible — or if the landowners, a
large part of whom presumably were farmers themselves, communally agreed to
support their church in this way. In either case common interests are evident.
It is of course a tautology to claim that a parish represents a community. It
clearly does so just as larger units — for example, communes (associations of
twenty farms or more, normally larger than parishes and often not coterminous
with them34) or assembly areas (areas of one hundred farms or more attending the
same judicial court) — constituted communities in some sense. It does not
necessarily follow that these communities were significant for people’s identities
or everyday behaviour. In the case of the Icelandic parishes, however, they clearly
were, to a different degree. Some parishes were densely populated, with a number
of other factors than a common church tying people together; others were much
31
Diplomatarium Islandicum, IV , 95–96.
32
Vésteinsson, McGovern and Keller, ‘Enduring Impacts’, pp.121–23.
33
Diplomatarium Islandicum, III, 239.
34
See Callow, this volume.
Page 124
Ordernr. 061297 P1
more dispersed and had few or no other factors helping to create the conditions
for community feeling. The basic structure of Icelandic settlements, however,
betrays their conceptual origin as small and tightly-knit communities.
I have argued elsewhere that Icelandic settlements can be divided into two:
clustered settlements — which come in at least three different types — and
planned settlements.35 The former seem to be primary settlements, those of the
greatest antiquity, whereas the latter seem to be secondary, planned and settled
after the former type of settlement had been established, presumably in the late
ninth century. Planned settlements are by definition modest, single household
holdings, most often quite dispersed, of which the Öxnadalur map shows a good
example (Fig. 5.1). They represent more than half of all Icelandic farms. When
their arrangement into parishes is studied it becomes clear that the planned
settlements in each parish occupy the immediate periphery of a clustered settle-
ment, the postulated primary settlement. Some parishes are made up of a single
such set (i.e. a cluster plus one or more planned settlements) while others are more
complex and may contain remains of more than one set. The church at Núpar
mentioned above seems to indicate that the farm originally constituted a part of
such a set that later became subsumed in the parish of Reykir.
The reality of such sets can be argued from the endowments of churches that
seem mostly to have taken place in the twelfth century.36 An example can be taken
of Laugardælir in Flói (Fig. 5.4). According to its earliest charter the church there
owned the whole estate as well as two adjacent farms, Svarfhóll and Uppsalir.37
They may not have been parts of the original endowment but Svarfhóll, judging
from its location and place-name, is likely to have been a part of the original estate.
Uppsalir on the other hand is one among many in a strip of modest farms ringing
the estate of Laugardælir on the south. The place-name indicates a place upwards
from some other place — which in this case can only be Laugardælir as the adja-
cent farms to Uppsalir are situated on the same slight rise, in an otherwise rather
flat landscape. It is arguable then that Uppsalir also belonged to the original estate
and that when the owner of Laugardælir decided to endow his church he decided
to donate the core of the estate, the central holding, and two of the more
peripheral ones, while presumably retaining the rest for his and his heirs’ private
35
Vésteinsson, ‘Patterns of Settlement’, pp.17–23; Vésteinsson, McGovern and Keller,
‘Enduring Impacts’, pp.117–25.
36
Vésteinsson, Christianization of Iceland, pp.101–09.
37
Diplomatarium Islandicum, IV , 53.
Page 125
Ordernr. 061297 P1
Figure 5.4. Laugardælir parish.
use — farms that in the course of time became sold or inherited in different
directions. The reverse took place in Bær in Borgarfjörður in the 1180s where a
landowner decided to keep the core of the estate for himself and his heirs but
donated three outlying farms to his church.38
38
Vésteinsson, Christianization of Iceland, pp.112–13.
Page 126
Ordernr. 061297 P1
In the case of Laugardælir the parish is made up of three sets, the largest and
most complex of which is Laugardælir. The estate of Laugardælir seems originally
to have consisted of the central farm itself, the church farm Svarfhóll and on
average three cottages, Reykjavellir, Þorleifskot and Músarhóll, making up a
relatively tight linear settlement. Ringing the estate core are Ósgerði to the north
— a farm with unstable occupancy considered as a part of the estate in later cen-
turies — and Tún, Uppsalir and Sölfholt to the south. Of these Sölfholt was an
above average farm in terms of taxation value and had two cottages in 1703 —
possibly representing a very early division from the original estate. At least two
other farms, Lækur and Bár, both quite modest, are located on a still outer ring.39
The parish had thirteen farms40 and five of these were west of Sölfholt, but in a
different commune. Commune boundaries often dissect parishes, especially in
open landscapes like this one. This seems to indicate that the farms belonging to
a different commune from the estate where their parish church is represent
separate sets. They represent communities that could thus retain some of their
independence vis-à-vis their church lord on the one hand and the leader of their
commune on the other (who was most likely the church lord of the adjacent
parish). Apart from the one farm by the river, Selfoss, the rest of these make up a
linear settlement, quite dense in places, centred on two farms, Smjördalir and
Votmúli. On the northern side of Laugardælir there is the third and smallest set.
This is made up of three farms, Langholt with a church and cemetery,41 and two
farms at Ámót, with a total of six households in 1703.
Conclusions
The parish of Laugardælir exhibits all the basic elements of the Icelandic settle-
ment structure. It is made up of three neighbourhoods, the central one by far the
largest and presumably the earliest and always the most significant in economic
and political terms. Belonging to this central neighbourhood is a double line of
planned settlements, extending outwards from the estate. On either side of this
there are different types of neighbourhoods. On the western side there is a fairly
39
Compare Árnessýsla: Sýslu og sóknalýsingar Hins íslenzka bókmenntafélags 1839–1843 og
Lýsing Ölfushrepps anno 1703 eftir Hálfdan Jónsson, ed. S. Sigmundsson (Reykjavík, 1979),
pp.72–73.
40
AM 66a 8vo , pp.35a–37b.
41
AM 66a 8vo , pp.35a–37b.
Page 127
Ordernr. 061297 P1
dense scatter of modest holdings that have all the hallmarks of planned settle-
ments except that they belong to a different commune from their church lord,
indicating a degree of independence and separateness from the church estate. On
the northern side of Laugardælir estate there is a much smaller estate with its own
church in Langholt and planned settlements in Ámót.
The principal conclusions we can draw from this study are firstly that Icelandic
settlements were isolated only in a relative sense. In fact the actual distances
between settlements were not very much greater than elsewhere in northern
Europe. Rather, the settlements were smaller, with a much higher proportion of
single household settlements. Secondly — and more importantly — we can con-
clude that Icelandic settlements had a structured distribution and that this
structure indicates a preference for ordering settlements in such a way that they
made up neighbourhoods that, it is argued here, functioned as communities. This
suggests that the ties that bound people together and gave them a common
identity outweighed the physical distances that separated their settlements. In
some cases it seems that despite often-considerable distances between settlements
the sense of community was so strong that a formal name only existed for the
community, not the individual settlements. This can only have applied to the first
generations of settlers and may reflect their original organization, presumably
based on ties of kinship and dependency, cooperation and common enemies. At
some later date — possibly as early as the tenth century, in other cases maybe not
until late medieval times — these communities experienced fission and each
settlement got its own formal name and status as a separate property, in effect
signalling a loosening of the earlier community ties. Despite this fissioning, the
locations of the settlements remained in the vast majority of cases, providing the
spatial context for the inhabitants’ interrelationships for centuries to come.
Following the introduction of Christianity in the beginning of the eleventh
century, a parish system developed. There are no indications that this system was
imposed from above. Rather, it grew out of and reflected the settlement organi-
zation of the eleventh and twelfth centuries. It is important to stress that this
organization may have been significantly different from that of the settlement
period of the ninth to tenth centuries. In particular it may be that small commu-
nities evidenced by the place-names (based on kinship and colonizing coopera-
tion?) had ceased to be of primary significance and had been superseded by slightly
larger communities based more on ties of dependency and tenure. It is argued
here, however, that the basic characteristic of early Icelandic settlement was that
it was organized into communities, definable neighbourhoods that reflect the
shared identity and concerns of their inhabitants. On the basis of the evidence
Page 128
considered here it is, however, not possible to do more than guess at the nature of
these communities, or their importance for their members.
What can be suggested is that from the outset there must have been a very real
difference between the communities centred on a great estate with rings of
dependent households around them and communities that were made up mainly
of planned settlements. As a rule the former type of community had higher
population density but greater social and economic differentiation among the
inhabitants. In the second type of community the distances between settlements
were greater but the households were more alike socially and economically, and
may have had shared interests in relation to a common landowner and/or church
lord residing outside the community.
I have been concerned here to show that in its structure Icelandic settlement
organization betrays its origins as a series of communities, each comprising several
farmsteads and households, rather than isolated and independent farmsteads as
has been the accepted view. This conclusion has implications for our understand-
ing of the social organization of the settlers of Iceland. It suggests that they were
organized into groups of households and that these groups formed the building
blocks of the new society, rather than the single family households led by a free
farmer normally identified as the basic social unit. This makes the Icelanders less
of an anomaly among medieval European societies, and should — I hope — clear
the path towards a fuller understanding of the system of government that emerged
in Iceland in the course of the tenth and eleventh centuries.
As a general lesson from this I hope to have shown that appearances can be
deceptive: dispersed settlement does not necessarily mean a more diffuse social
organization. In fact I think the opposite could be argued — but this must await
another opportunity.
Page 129
Ordernr. 061297 P1
Page 130
Ordernr. 061297 P1
B OUNDARIES OF K NOWLEDGE:
M APPING THE L AND U NITS
OF L ATE A NGLO -S AXON AND N ORMAN E NGLAND
Steven Bassett
T
here is an essential spatial dimension to the continuing debate about the
origins of the English parochial system, which led to people’s pastoral care
being increasingly provided by a local church. It had not always been so.
In the late Anglo-Saxon period very large parishes*, each one served by a mother-
church*, are widely held to have existed throughout much, perhaps most, of
England; but the question of when they emerged — whether it was in the seventh
and early eighth centuries, or not until the tenth century or even later — is still
being strenuously disputed. We need to ask, therefore, how and how reliably we
can discover the extents of these mother-church parishes, date their origins and
grasp the historical significance of their individual shapes and sizes. Secondly, how
well can we understand the processes, and chart the stages, by which local com-
munities encouraged and benefited from the slow fission of most of the mother-
church parishes — a trend which by 1200 had seen the birth of the familiar
parochial geography of recent centuries set fully in motion?1
1
For the main substance of the debate, see C. N. L. Brooke, ‘Rural Ecclesiastical Institutions
in England: The Search for their Origins’, Settimane di studio del centro italiano di studi sull’alto
medioevo, 28 (1982), 685–711; J. Blair, ‘Introduction: From Minster to Parish Church’, in
Minsters and Parish Churches: The Local Church in Transition, 950–1200, ed. J. Blair, Oxford
University Committee for Archaeology, monograph 17 (Oxford, 1988), pp.1–19; E. Cambridge
and D. Rollason, ‘The Pastoral Organization of the Anglo-Saxon Church: A Review of the “Min-
ster Hypothesis”’, Early Medieval Europe, 4 (1995), 87–104; J. Blair, ‘Ecclesiastical Organization
and Pastoral Care in Anglo-Saxon England’, Early Medieval Europe, 4 (1995), 193–212;
D. Rollason, ‘Monasteries and Society in Early Medieval Northumbria’, in Monasteries and Society
Page 131
Ordernr. 061297 P1
116 Steven Bassett
When this spatial dimension has been addressed previously in discussions of

the origins of the parochial system, the results have often been invalidated by, for
instance, a reliance on late nineteenth- or even twentieth-century maps or the lack
of a sufficiently rigorous chronological framing.2 However, it has usually been
overlooked altogether, being considered, one suspects, too unimportant, too
unprofitable, and/or too unreliable to justify the large amount of research time
needed for its proper study. This is regrettable since it is none of these things. Few
historically significant aspects of early medieval life were wholly unaffected by
constraints imposed by the formal allocation of space. Individuals may normally
have had rather less awareness of the territorial dimensions of the several commun-
ities to which they contemporaneously belonged than of the nature of the shared
benefits and responsibilities which united each one’s members, but that does not
mean that none existed.3 In most areas of life fixed visible boundaries identified
the territory of each community to which an individual belonged — whether it
was a political and administrative, a socio-economic or an ecclesiastical one — and
were an essential physical statement of its members’ identity and common interests.
Many Anglo-Saxon and even earlier boundaries, reflecting decisions dictated
by a group’s or individual’s relations with the natural and human landscape, with
neighbours, and with their rulers, have survived into the age of accurate map-
making and air-photography, sometimes as relict features but frequently by con-
tinuing in use. Time spent in trying to map the various overlapping territories
within which specific communities operated at a given period, and in under-
standing what determined each one’s extent and how it related physically and
institutionally to the others, is therefore rarely wasted. Moreover, when the
available evidence allows firm conclusions to be drawn, the rural land units of
in Medieval Britain, ed. B. Thompson (Stamford, 1999), pp.59–74; and now J. Blair, The Church
in Anglo-Saxon Society (Oxford, 2005), passim.
2
A great many changes were made to English parish boundaries in the course of the nine-
teenth century, as a result of which almost no Ordnance Survey map can be safely relied on to
show unchanged ones (including, despite its title, the ‘Index to the Tithe Survey’ edition of the
Old Series one-inch maps). Only parliamentary enclosure maps and tithe maps (where they are
available), most of which pre-date the changes, are a generally reliable source for ‘ancient’ parish
boundaries. For a convenient summary of these changes, see R . J. P. Kain and R . R . Oliver, The
Historic Parishes of England and Wales (Colchester, 2001), pp.7–8. For an example of the value
of assiduous data collection and map work being seriously undermined by poor attention to
chronological framing, see T. A. Hall, Minster Churches in the Dorset Landscape, British
Archaeological Reports, British Series, 304 (Oxford, 2000).
3
In some contexts, for example in the conduct of commerce, the dimensions were too
informal, extended and fluid for the setting-out of lines on the ground to be either useful or viable.
Page 132
Ordernr. 061297 P1
BOUNDARIES OF KNOWLEDGE 117
medieval England characteristically show an intimate physical association with

one another; and the Church’s boundaries, those of landholding and farming, and
those of secular administration and justice are found to have coincided, either
one-to-one or with one land unit being coterminous with several of another sort.4
Therefore, since manors are widely accepted as having been the building-blocks
of parishes and hundreds* alike,5 mapping and analyzing the changing geography
of the medieval English Church throws light on secular communities and institu-
tions as well as on ecclesiastical ones.
There has been a great deal of debate during the last twenty-five years about
the origins of the English parochial system. Everyone agrees that by 1200 most
parishes already existed in the form which they would then maintain with few
changes until the middle years of the nineteenth century, and in some cases right
up to today. There is also general agreement about the later stages of the processes
which produced this late medieval and modern parochial geography. By the late
Anglo-Saxon period, immediately below the cathedrals were many important
churches spread across England which, then as now, were called mother-churches.
Each had a parish which was normally far larger than the parishes of the late
medieval and modern periods;6 and almost all the churches of this sort about
which we know in detail were already many centuries old by then, having been
founded in the missionary phase of Anglo-Saxon Christianity, that is, in the
seventh and early eighth centuries. These are often called ‘old minsters*’ in
modern secondary sources.7
4
For example, P. H. Hase, ‘The Mother Churches of Hampshire’, in Minsters and Parish
Churches, ed. Blair, pp.45–66; S. Bassett, ‘The Administrative Landscape of the Diocese of
Worcester in the Tenth Century’, in St Oswald of Worcester: Life and Influence, ed. N. Brooks and
C. Cubitt (Leicester, 1996), pp.147–73 (pp.160–68).
5
For example, F. W. Maitland, Domesday Book and Beyond (Cambridge, 1897), p.15 and
passim; R . Lennard, Rural England 1086–1135 (Oxford, 1959), pp.5–6, 9, 17. For detailed exem-
plification: C. D. Drew, ‘The Manors of the Iwerne Valley, Dorset’, Proceedings of the Dorset
Natural History and Antiquarian Field Club, 69 (1948), 45–50; S. Bassett, ‘Continuity and
Fission in the Anglo-Saxon Landscape: The Origins of the Rodings (Essex)’, Landscape History,
19 (1997), 25–42.
6
I use the word parish as defined in the glossary. This usage applies as much to what I refer
to here as ‘mother-church parishes’ as it does to the many much smaller parishes produced by their
fission. The nature of available pastoral care is likely to have altered considerably between the
seventh century and the twelfth, but not to so great an extent that my use of the term ‘parish’ is
self-evidently inappropriate for any part of this period.
7
This echoes the term used in Edgar’s second law code: Die Gesetze der Angelsachsen, ed.
F. Liebermann (Halle, 1903), p.196 (II Edgar, §1.1, 2.2).
Page 133
Ordernr. 061297 P1
118 Steven Bassett
It is also widely agreed that by the late twelfth century two other, less impor-
tant sorts of rural church had come to prominence, many of which succeeded in
carving out parts of the mother-churches’ parishes to form their own much
smaller independent ones. One sort were chapels* built in appropriate outlying
parts of a mother-church parish and staffed by one or more priests. They were
public churches from the outset, set up to help the mother-churches to deliver
pastoral care to all parts of their extensive parishes. Each of these parochial
chapels* was at first fully subordinate to its mother-church, but in many cases it
sooner or later gained independent control of the often large part of the parish
which it had been founded to help serve.8
Lesser rural churches of the other sort were built in large numbers by laymen
in the tenth to twelfth centuries, almost always next to a manor-house. Their
initial role may simply have been to offer a suitable venue for the private devotions
of the local lord and his household; but in time many manorial chapels gained a
public role through being consecrated by the bishop, given a priest, and allowed
a measure of pastoral responsibility under the control of the mother-church or
parochial chapel in whose area of parochial authority they stood. By the late
eleventh century bishops were actively encouraging local lords to build churches,
and were assiduous in consecrating them so as to help spread the pastoral load.9
Eventually, a great many of these manorial chapels managed to break free and
become effectively independent parish churches, each one serving as its parish the
manor on which it stood.
By 1200 the parochial system of late medieval and modern England had
evolved, with only a small number of mother-churches still retaining control of
the entirety of their original parish. Most of it was normally being served by the
lesser churches, with some of them already independent but others merely begin-
ning to gain parochial rights and not breaking entirely free for decades or even
centuries to come. In many areas, therefore, a three-level hierarchy of churches
was involved in delivering pastoral care at the sub-diocesan level, with some man-
8
Compare P. H. Hase, ‘The Church in the Wessex Heartlands’, in The Medieval Landscape
of Wessex, ed. M. Aston and C. Lewis, Oxbow Monograph, 46 (Oxford, 1994), pp.47–81 (p.63:
‘dependent parochial chapelry’).
9
For example, Wulfstan II of Worcester: The Vita Wulfstani of William of Malmesbury, ed.
R . R . Darlington (London, 1928), pp.32, 45, 52 (‘Per totam parrochiam in sui iuris prediis
ecclesias struebat; in alienis ut struerentur instabat’; ‘He [Wulfstan] built churches throughout
the diocese on his own lands; on other people’s he insisted that they should be built’), 54–55
(respectively, II.9, II.22, III.10, III.14–15).
Page 134
Ordernr. 061297 P1
orial chapels being directly subordinate to a mother-church but others only indir-
ectly so because they had been set up in the pastoral area of a parochial chapel.
There is general agreement about these late Anglo-Saxon and later processes
in the development of the English parochial system. Where there is disagreement,
and it is a strongly entrenched one, it is about the origins of mother-churches and
their very large parishes. Even here most scholars agree that ‘old minsters’ were
characteristically founded, diocese-by-diocese, in the seventh and early eighth
centuries and endowed by the newly converted Anglo-Saxon kings. We know
about some of them from the charters which were issued either at the time of their
foundation or at a later date. Each housed a more or, usually, less genuinely
monastic community and also a number of priests who ministered to its mem-
bers.10 Most of these churches still exist as parish churches but are now largely
indistinguishable from the very many churches of later, more lowly origin.
The academic disagreement is focussed on when the ‘old minsters’ gained a
pastoral role — whether from the very start, so that what we see happening in the
tenth to twelfth centuries was the transformation of an old, badly failing system;
or, alternatively, not until the tenth century or afterwards, prior to which they had
had no role in delivering pastoral care. The tenth and eleventh centuries, then, lie
at the heart of the battleground. To some scholars this was when the long-
established ‘minster system’ began to be supplanted by a new system based on local
churches which typically served either a single manor or no more than a cluster of
two or three. To others it was when ‘old minsters’ first gained parishes: England
was being politically unified — a lengthy and unevenly successful process — and
its kings were striving to impose some uniformity on it, with one of their measures
being to have a parish created around each already old ‘old minster’ and to give its
priests a new, external, pastoral role.11
A Local Study: The Original Parish Served by St Peter’s, Wootton Wawen
I shall not engage with this debate until the final part of this essay. My primary
aim is to demonstrate how we can find out the extents of the large mother-church
parishes which are widely held to have been in existence by the late Anglo-Saxon
10
Cambridge and Rollason, ‘Pastoral Organization’, pp.90–91, 94–95; Blair, ‘Ecclesiastical
Organization’, pp.208–09.
11
Blair, ‘Introduction’, pp.7–9; Cambridge and Rollason, ‘Pastoral Organization’, pp.97–103;
Blair, ‘Ecclesiastical Organization’, pp.197–98.
Page 135
Ordernr. 061297 P1
120 Steven Bassett
period — irrespective of whether they were already very old by then or had only
recently been created. This is hardly ever done: the existence of mother-church
parishes at this period has usually been viewed as uncontroversial, but they have
only rarely been mapped by detailed local studies.12
Specifically, I want to explain how we can discover the extent of the original
parish of the church of Wootton Wawen (Warwicks.), which is in the west
midlands c.25 km south of Birmingham. It is a church with many structural
phases, the first of which is of tenth- or early eleventh-century date and includes
a central tower.13 It replaced a church which had been built in the early eighth
century, and for which the foundation charter survives in an early eleventh-
century hand in a Worcester cartulary*.14 Issued by the Mercian king Æthelbald
between 716 and 737, it speaks of a minster being set up at Wootton and
endowed with land, assessed at twenty hides*, which lay within a territory (regio)
called Stoppingas. This now lost name, the equivalent of ones such as Hastings,
Reading and Roding, means ‘people of the bucket-shaped hollow’ or ‘Stoppa’s
people’, in the sense of ‘the land of the people of the bucket-shaped hollow’ or ‘the
land of Stoppa’s people’.15
Wootton Wawen’s church is, then, undoubtedly a minster of early foundation.
Today its parish covers 3187 ha, which is large by west midland standards (Fig.
6.1); but it is only the residue of what the church served in the late Anglo-Saxon
12
Examples are D. A. Hinton and C. J. Webster, ‘Excavations at the Church of St Martin,
Wareham, 1985–86, and “Minsters” in South-east Dorset’, Proceedings of the Dorset Natural
History and Archaeological Society, 109 (1987), 47–54 (pp.50–53); Hase, ‘Church in the Wessex
Heartlands’; S. Bassett, The Origins of the Parishes of the Deerhurst Area (Deerhurst Lecture 1997,
1998); D. M. Palliser, ‘The “Minster Hypothesis”: A Case Study’, Early Medieval Europe, 5
(1996), 207–14 (about Beverley, but regrettably not illustrated with a map).
13
H. M. Taylor and J. Taylor, Anglo-Saxon Architecture, 2 vols (Cambridge, 1965), II,
685–88; E. Fernie, The Architecture of the Anglo-Saxons (London, 1983), pp.116, 128, 178;
S. Bassett, The Wootton Wawen Project: Interim Report No. 1 (University of Birmingham, School
of History, 1983), pp.2–9; Interim Report No. 2 (1984), pp.2–6; Interim Report No. 3 (1985),
pp.1–14; Interim Report No. 4 (1986), pp.1–12; Interim Report No. 5 (1987), pp.1–8; Interim
Report No. 6 (1988), pp.1–9.
14
Cartularium Saxonicum, ed. W. de G. Birch, 3 vols (London, 1885–93), I, no. 157; P. H.
Sawyer, Anglo-Saxon Charters: An Annotated Hand List and Bibliography (London, 1968), no. 94.
15
M. Gelling, ‘The Place-Name Volumes for Worcestershire and Warwickshire: A New
Look’, in Field and Forest: An Historical Geography of Warwickshire and Worcestershire, ed. T. R .
Slater and P. J. Jarvis (Norwich, 1982), pp.59–78 (p.69).
Page 136
Page 137
Ordernr. 061297 P1
BOUNDARIES OF KNOWLEDGE
Figure 6.1. Parishes in the vicinity of Wootton Wawen (Warwicks.) at their late eighteenth-/early nineteenth-century extent.
Oldberrow was transferred into Warwickshire in 1896.
121
Bushwood was a detached part of the parish of Old Stratford (not on the map).
Ordernr. 061297 P1
122 Steven Bassett
period when its parish arguably covered c.7000 ha (Fig. 6.2).16 To prove this state-
ment is, however, no easy matter since it involves assembling scraps of evidence
from a wide range of sources — mostly from documents, but some from place-
names and a few more from maps drawn in the last 250 years. When assembled,
this evidence indicates that as many as eleven other churches were once subordi-
nate to Wootton Wawen’s church. Most of them sooner or later broke free of its
control, and today they serve the other parishes shown on Fig. 6.1.17 To rephrase
this in appropriate medieval terms, Wootton Wawen’s church was once the
mother-church of the whole area depicted there, and the eleven other churches
were at first no more than subordinate chapels which stood in different parts of
its originally far larger parish. Ten of the eleven still exist and have much late
medieval fabric in them but, unlike Wootton Wawen’s church, none has anything
Anglo-Saxon.
Henley-in-Arden
Of these churches the latest to be founded was St John’s, Henley-in-Arden. Until

1914 it was subject to the church of Wootton Wawen.18 Henley is a small, single-
street borough created in the twelfth century, or soon after, by being laid out with a
regular plan on a greenfield site at the south-eastern corner of the manor of Ullenhall.
In 1367 the bishop of Worcester let the burgesses build a chapel dedicated to St
John, so long as they acknowledged that they were still subject to Wootton Wawen’s
church.19 The latter had successfully petitioned for a chapel of their own for the
usual reason — that the journey from where they lived to their mother-church
16
Respectively, 7875 acres and nearly 17,000 acres in size.
17
Fig. 6.1 is based on the following maps held at Warwick County Record Office: DR
195/28 (Wootton Wawen: 1736); Z 190/6(U) (Wootton Wawen, Bearley and Beaudesert:
1800); CR 569/278 (Wootton Wawen, incl. Ullenhall and Henley: 1843); QS 75/79 (Morton
Bagot: 1807); CR 1094/1 (Morton Bagot: 1821); CR 569/188 (Oldberrow: 1839); CR 569/23
(Beaudesert: 1841); CR 569/192 (Preston Bagot: 1840); CR 569/65 (Claverdon, incl. Langley,
and Norton Lindsey: 1841); QS 75/134 (Wolverton: 1826); CR 569/276 (Wolverton: 1849);
CR 611/607 (Bearley: c.1775).
18
VCH, Warwickshire, vol. III, ed. P Styles (London, 1945), p.211.
19
Worcester County Record Office, BA 2648/4 (ii), ref. no. b716.092, fol. 17 r (printed in
W. Dugdale, Antiquities of Warwickshire, 2nd edn in 2 vols (London, 1730), II, 806); D. Graham,
‘The Foundation Charter of the Chapel of St John the Evangelist at Henley-in-Arden, 1367’,
Warwickshire History, 12 (2003–04), 154–68.
Page 138
Page 139
Ordernr. 061297 P1
Figure 6.2. Parishes, chapelries, hamlets, and other significant land units in the vicinity of Wootton Wawen (Warwicks.) at
their late eighteenth-/early nineteenth-century extent. (The boundary shown between Offord and Wootton Wawen is only
123
approximate.) The probable maximum extent of the parish served by St Peter’s Wootton Wawen is also shown.
Ordernr. 061297 P1
124 Steven Bassett
was too hard to make in winter because of distance and the awful state of the
roads. In reality, they were undoubtedly motivated rather more by wanting a mea-
sure of ecclesiastical independence, as befitted their burghal status, than by the
rigours of the short gentle journey from Henley to Wootton Wawen. Although
taking corpses to a mother-church for burial in winter was often a serious problem
for the people of outlying parts of a big parish, Henley’s church has never had a
graveyard. Instead, specific areas of Wootton Wawen’s graveyard seem to have
been exclusively reserved for the dead of Henley, who were being buried there
until 1915.20 In 1914 St John’s at last became independent and was given the area
of the borough as its parish; but less than a year later it lost its new-found freedom
when the living was combined with that of neighbouring Beaudesert (whose
church lies only 200 m to the east, on the opposite bank of the River Alne).21
Ullenhall
Ullenhall had a church by the twelfth century, but the whole area was subject to
Wootton Wawen’s church for almost as long as Henley-in-Arden was, not becom-
ing an independent parish until 1861.22 Structural evidence suggests that the
church existed by the twelfth century, but it is not recorded until 1586, when an
ecclesiastical survey states that it was entitled to have a priest supplied by the vicar
of Wootton Wawen at his own cost.23 The church has a fifteenth-century font,
and both the Wootton Wawen registers and the surviving memorials prove that
people were being buried there by the seventeenth century.24 Although it was only
a chapel, it may have had the right to baptize and bury throughout the late Middle
Ages. It is impossible to say when or how this church originated. Its present,
isolated site about 1 km north-east of Ullenhall village is striking. It stands over
20
I. A. Shorters, A Wasting Historical Asset? A Comparative Study of Grave Memorials at
Wootton Wawen, King’s Norton and Birmingham, c. 1700–1940, British Archaeological Reports,
British Series, 366 (Oxford, 2004), p.23.
21
VCH, Warwickshire, III, ed. Styles, pp.48, 211. The area shown as Henley’s parish on Fig.
6.2 is that of the township at its 1843 extent; the area of the late medieval borough was almost
certainly much smaller.
22
Ibid., p.215.
23
Minutes and Accounts of the Corporation of Stratford-upon-Avon and Other Records,
1553–1620, transcribed R . Savage and others, Publications of the Dugdale Society, 10 (London,
1929), IV , 3.
24
W. Cooper, Wootton Wawen: Its History and Records (Leeds, 1936), pp.123–25.
Page 140
Ordernr. 061297 P1
800 m from any other known medieval settlement, being located, however, at an
important medieval road junction.
These pieces of evidence all point in one direction. The church at Ullenhall
was probably founded by the priests of Wootton Wawen as a parochial chapel —
that is to say, founded to provide pastoral care to the population of a substantial
part of its mother-church’s parish. Many chapels of this sort already existed before
1066 and may often have replaced a preaching cross. They gained a font and a
graveyard early on, and in the west midlands at least their main characteristics are
the large area which each one served, often containing many Domesday manors,
their location on accessible sites in open countryside, and the poor quality of their
construction. In Ullenhall’s case the evidence of the 1586 survey (that its priest’s
salary was paid by the vicar of Wootton Wawen) and its inability to break free of
its mother-church until the late nineteenth century both reinforce this proposed
identification of it as, in origin, a parochial chapel.
Oldberrow
To the south-west of Ullenhall lies Oldberrow.25 By 1066 the manor of Oldber-

row belonged to Evesham Abbey. It did not lie in Warwickshire with all the rest
of Wootton Wawen’s original parish but was a detached part of Worcestershire,
belonging to the abbey’s private hundred of Fishborough.26 In the late Middle
Ages the monks had a charter which alleged that in 709, in return for a pound’s
measure of purest gold, the Mercian king Coenred had granted land at Oldberrow
to Bishop Ecgwine of Worcester for his lifetime, and that after his death it was to
pass to the abbey.27 The charter is a twelfth-century forgery, but it is of real
interest because it is accompanied in the cartulary by a boundary clause (in the
same hand) which begins, ‘These are the bounds of the land at Oldberrow which
25
Ullenhall’s Old English name means ‘nook of land of the owl’ and Oldberrow’s means ‘hill
of the owl’: J. E. B. Gover, A. Mawer and F. M. Stenton, The Place-Names of Warwickshire,
English Place-Name Society, 13 (Cambridge, 1936), p.245; A. Mawer and F. M. Stenton, The
Place-Names of Worcestershire, English Place-Name Society, 4 (Cambridge, 1927), pp.267–68;
M. Gelling and A. Cole, The Landscape of Place-Names (Stamford, 2000), pp.129, 150. (I am
grateful to Margaret Gelling for her further comments on these two names, pers. comm., 31
March 2002.) The shared first element of their names does not of itself show that the two areas
concerned had once belonged to a single larger territory.
26
DB, fol. 175b; DB Warwicks., 10,4.
27
Cartularium Saxonicum, ed. Birch, I, no. 124; Sawyer, Anglo-Saxon Charters, no. 79.
Page 141
Ordernr. 061297 P1
126 Steven Bassett
King Beorhtwulf freed and booked to Evesham’ (Beorhtwulf having been king of
Mercia 840–52).28 No such privilege is otherwise known about; but, even so,
Beorhtwulf’s alleged role in the history of Evesham’s possession of Oldberrow has
the ring of truth. As one of the less well-known kings of Mercia he is unlikely to
have been chosen by a twelfth-century scribe who was intent on forging a convinc-
ing account of how Evesham gained Oldberrow. This would be true even if no
charter of the early eighth-century king Coenred had been forged at the same
time. However, it was forged then; and so there was nothing to be gained by
spuriously attributing anything to Beorhtwulf. I deduce, therefore, that there was
an apparently authentic tradition at Evesham that it had held Oldberrow in
Beorhtwulf’s time, that is, in the mid-ninth century; but whether he had been the
original donor or, as is more likely, had merely freed land which Evesham already
held from most of its fiscal burdens is unknowable.
Despite Evesham Abbey’s apparently long association with the place, Oldber-
row’s church was required to recognize Wootton Wawen as its mother in the late
twelfth century. In c.1150 the bishop of Worcester let the chapel of Oldberrow
have a graveyard. But soon a dispute about burial rights arose between Evesham
and the Norman abbey of Conches, which had been holding Wootton Wawen’s
church since the late eleventh century. It was settled in favour of Conches, as a
result of which the priest of Oldberrow thereafter had to pay two shillings each
year to Wootton Wawen in place of the individual burial fees.29 This is unim-
peachable evidence that Oldberrow lay in its parish.
The chapel itself is an essentially twelfth-century building. Its location next to
the manor-house and its close association with the Dispenser family, who were
Evesham’s tenants at Oldberrow in the twelfth century,30 together suggest that it
was of manorial origin. Never a structure of much architectural distinction, it is
28
For a suggested solution of the boundary clause, see D. Hooke, Warwickshire Anglo-Saxon
Charter-Bounds (Woodbridge, 1999), pp.30–36.
29
Warwicks. Record Office, CRO MI 331, charter 87 (translated in Cooper, Wootton
Wawen, p.42). The payment was discontinued by a final judgement of c.1186, made at the request
of Pope Urban III, which nonetheless fully confirmed the matronal rights of Wootton Wawen’s
church in Oldberrow: Warwicks. Record Office, CRO MI 331, charter 89. Conches (alias St
Peter of Castellion) had received the church of Wootton Wawen from its Domesday holder,
Robert of Stafford (d. 1088): ‘Instrumenta ecclesiae Ebroicensis’, no. V, in ‘Instrumenta ad
tomum xi’, in Gallia Christiana in Provincias Ecclesiasticas Distributa, vol. XI, ed. P. Piolin (Paris,
1874), cols 128–33 (col. 131: ‘ecclesiam de Octona’).
30
VCH, Warwickshire, III, ed. Styles, p.140.
Page 142
Ordernr. 061297 P1
still largely in its twelfth-century form. When the chapel was allowed to have a
graveyard of its own, Conches recognized a serious threat to its income from
Wootton Wawen’s church and succeeded in enforcing the right of Oldberrow’s
mother-church to the payment owed to it whenever any of the manor’s popula-
tion were buried there.
What is especially interesting about this particular case is that Wootton
Wawen was entitled to the burial fees of a manor which had, it seems, been in
Evesham Abbey’s hands since at least the mid-ninth century. In many other cases
we find that the burial fees and other dues of manors like Oldberrow did end up
being paid to a distant monastic lord such as Evesham. In studies of other areas I
have found that it was not unusual for a manor which was granted to a large
monastic church elsewhere in the same region to be ecclesiastically ‘captured’
sooner or later by the latter.31 This was presumably the result of the manor’s real
mother-church failing to make a fuss about developments such as those which
occurred at Oldberrow in the mid-twelfth century, or perhaps even failing to
notice them. However, in Oldberrow’s case the mother-church’s owner did notice
and took swift action — a victory which throws important light on the date of
creation of Wootton Wawen’s original parish (below, pp.137–39).
Morton Bagot
Immediately south of Oldberrow lies Morton Bagot. Its church too is next to the
site of the former manor-house and also seems to have originated as a manorial
chapel. It stands just outside a splendidly preserved ringwork*, which contained
the manor-house throughout the late medieval period. The present church’s nave
still has one round-headed splayed window, and its thirteenth-century chancel
arch sits on cushion capitals, which together indicate that the structure was in ex-
istence by the twelfth century. Morton Bagot lies west of the watershed separating
the valley of the River Alne from that of its western neighbour the Arrow, and so
in geographical terms it does not belong to Wootton Wawen’s natural hinterland.
There can be no doubt, however, that it once belonged to its parish. Early in the
fourteenth century there was a dispute about the burial fees which had to be paid
whenever an inhabitant of Morton Bagot died. By then its church had been in the
hands of Kenilworth Priory for over a century and a half as the result of a grant
31
S. Bassett, ‘Anglo-Saxon Birmingham’, Midland History, 25 (2000), 1–27 (pp.10–12);
S. Bassett, Anglo-Saxon Coventry and its Churches, Dugdale Society Occasional Papers, 41 (2001),
pp.22–23.
Page 143
Ordernr. 061297 P1
128 Steven Bassett
made c.1253x54 by the manor’s lord, Peter de Mora.32 When the dispute arose,
Kenilworth said that it had a right to three-quarters of the payment and that only
a quarter of it should go to Wootton Wawen. But the latter’s right to the full
amount was finally upheld by an order in Chancery in 1325x26.33
This corroborates what we can safely deduce from the location of Morton Bagot’s
church next to the site of the manor-house and from Peter de Mora’s ability to give
it to Kenilworth Priory in the 1250s — that is to say, that it had begun as a manorial
chapel in an area which was parochially subject to Wootton Wawen’s church. By
the early fourteenth century the chapel was largely independent of Wootton Wawen,
having its own graveyard by then (as the dispute about burial fees shows). It took
Conches far longer to insist on the rights of Wootton Wawen’s church at Morton
Bagot, but when it did the result was the same as at Oldberrow. Not withstanding
the firm grip that the distant Kenilworth Priory had on Morton Bagot’s church
and its land, the real mother-church won the dispute and in doing so it provided
us with clear evidence that Morton Bagot had once lain in its parish.
Bearley
The parish of Bearley presents a different situation. In 1220x21 John of Bearley

gave Wootton Wawen’s church his share in the advowson of Bearley chapel.34
Superficially, this suggests that until then the chapel had been subject to some
other mother-church, but that is not so. More than thirty years before, in 1189,
John’s father William had given his rights in Bearley’s chapel to Bordesley Abbey
(near Redditch in Worcestershire), along with a virgate of land. John himself
confirmed this gift after his father’s death;35 but he evidently changed his mind,
and as a result he dispossessed Bordesley Abbey in 1220x21 and instead gave his
rights in it to the priory which Conches had founded in the church at Wootton
Wawen. It is hard to be certain what this means, but the best explanation is that
Wootton Wawen, or probably Conches, had belatedly but successfully opposed
William of Bearley’s decision to give away his rights in the chapel. If so, it was
presumably on the grounds of its being a chapel which was anciently subject to
Wootton Wawen’s church.
32
Dugdale, Antiquities, II, 760.
33
Ibid.
34
Warwickshire Feet of Fines, vol. I, ed. F. C. Wellstood, Publications of the Dugdale Society,
11 (London, 1935), p.57, no. 294.
35
Page 144
Ordernr. 061297 P1
There were two manors of Bearley in the late twelfth century, one of them held
by William of Bearley and the other one being in Bordesley’s hands.36 The chapel
itself almost certainly stood on the latter, but it is clear that rights in it were also
vested in William’s manor. These presumably dated from a time when there was only
one manor of Bearley, so that when it had been divided into two at some unknown
date before 1066, rights in the chapel had been similarly divided.37 William’s grant
of his rights in it to Bordesley Abbey allowed the latter’s monks to exercise full
control of the chapel and no doubt to get their hands on all its income. Now, if this
chapel had been of manorial origin, Conches would arguably have had no grounds
for opposing what William had done, and so it is reasonable to deduce that in origin
it was a parochial chapel. By the late eleventh century most, perhaps all, of the
churches of this sort had become the responsibility of the lords of the manors on
which they stood, which explains why confusion could arise at a later date about
who exactly had rights in a parochial chapel. Where the mother-church failed to
insist on its own original rights, a church such as Bearley’s probably was might end
up as just another possession of the lord of the manor on which it stood.
This interpretation seems to be corroborated by later evidence. In 1249 John
of Bearley’s son William gave Bordesley Abbey the next ten years’ crops of his
demesne land* there, no doubt as compensation for what the abbey had lost in
1220x21.38 A list of parishes in the diocese of Worcester of 1563 reports that
Bearley’s church was a chapel of Wootton Wawen. By 1586 it had its own grave-
yard but paid a fee of 4d. for each burial to Wootton Wawen. It was still regarded
as Wootton Wawen’s chapel in 1786, and it remained a curacy* until in 1929 it
was united with Snitterfield, the next parish to the east.39
The church’s structure unfortunately adds very little evidence to what we have
already learnt. Because of several extensive alterations not much datable medieval
fabric remains. What there is, especially the north doorway, reveals that a small,
unicellular chapel of no special architectural merit existed in the twelfth century,
and that all the alterations made to it since then were both piecemeal and unpleas-
ant. That is exactly what I should expect in a chapel of such lowly status.
36
VCH, Warwickshire, III, ed. Styles, p.43.
37
DB, fols 242b, 243a; DB Warwicks., 22,27; 28,18.
38
T. Madox, Formulare Anglicanum (London, 1702), pp.134–35, no. 227.
39
London, British Library, MS Harley 595, fol. 212 v ; Dugdale, Antiquities, II, 831; J. Bacon,
Liber Regis, vel Thesaurus Rerum Ecclesiasticarum (London, 1786), p.990; VCH, Warwickshire,
III, ed. Styles, p.44.
Page 145
Ordernr. 061297 P1
130 Steven Bassett
Claverdon and its Chapels
In discussing Henley-in-Arden, Ullenhall, Oldberrow, Morton Bagot and Bearley

I have been dealing with churches for which hard evidence exists of their former
subordination to Wootton Wawen’s church. I come now, however, to five
churches for which there is no direct evidence, but which can nevertheless be
shown to have served original parts of its parish, together with a sixth one which
probably represents a ‘capture’ from the parish of an adjacent mother-church. It
is likely that in the Anglo-Saxon period Claverdon, not Wootton Wawen, was the
most important secular centre in the area once served by the latter’s church, and
that there was an Anglo-Saxon royal farm in Claverdon — the king’s tû n* to
which the area’s population were required to take the feorm*, their annual food
render. In the south-western part of Claverdon parish and the adjoining part of
Preston Bagot lay the Domesday manor of Kington. Perhaps it was so named
because there was a royal residence there, or perhaps it was not first called ‘the
king’s tûn’ until it was the only piece of royal land left in the entire area. By 1066,
however, it was no longer in the king’s hands.40
Claverdon’s importance even in ecclesiastical matters is shown by its having
a church by 1086, and by the fact that the latter had three chapels in the mid-
twelfth century. We might expect such a church to have been an original mother-
church itself, but there are good reasons for arguing that it had, instead, been set
up as yet another parochial chapel of Wootton Wawen’s church.41 In 1151x57 the
church of Claverdon was given by Hugh of Hatton to St Florent’s Abbey at
Saumur in Anjou, together with its three chapels of Langley, Norton Lindsey and
Wolverton.42 This is our first evidence of these chapels’ existence. (The earliest
datable fabric in the existing churches at Wolverton and Norton Lindsey is
thirteenth-century; and the chapel at Langley did not survive the Middle Ages,
with not even its site now being known.) Hugh’s grant to St Florent’s was, it
seems, abortive, but there is no need to doubt that it did occur and therefore that
40
DB, fol. 240a; DB Warwicks., 16,19. The evidence showing that part of the Domesday
manor of Kington lies in Preston Bagot parish is too lengthy to set out properly here; but in brief
the manor was subdivided after 1086, with one part being given to Bordesley Abbey, and with the
other part being administered together with the demesne manor of Preston Bagot and conse-
quently being parochially ‘captured’ by its church. (For ‘capture’ see below, p.133.)
41
DB, fol. 240a; DB Warwicks., 16,16. The church contains no pre-fifteenth-century fabric.
42
Calendar of Documents Preserved in France, vol. I, ed. J. H. Round (London, 1899), p.412,
no. 1146; p.414, no. 1148 (pt).
Page 146
Ordernr. 061297 P1
these three churches did exist in the 1150s and were believed, by some people
anyway, to be subject to Claverdon church. We do not know why Hugh’s grant
was annulled, but a likely explanation must be the intervention of Conches.
A dispute broke out c.1208 between Conches and the rector of Claverdon.43
This concerned the right of the latter’s church to take tithes and other ecclesias-
tical revenue from Langley and Norton Lindsey. (Wolverton was not involved
because by then it was already an independent parish church.)44 Conches claimed
that this income belonged by right to their church at Wootton Wawen, and it was
able to produce supporting documentation. The dispute dragged on for half a
century; but eventually in 1257 the chapels of Langley and Norton Lindsey and
their tithes were confirmed to Claverdon church, and the inhabitants were
instructed to take their dead to its yard for burial. In return Claverdon agreed to
pay Wootton Wawen each year the substantial sum of five marks and eight
shillings (£3 14s. 8d.).45
To modern eyes this was a triumph for common sense, even if it had taken fifty
years of litigation to achieve it. The very length of time needed is eloquent
testimony to the complexity of the dispute. At a hearing before papal delegates
c.1208 the rector of Claverdon had claimed that the ecclesiastical revenues of
Langley, Norton Lindsey and the whole of Wolverton had belonged to his church
‘by parochial right (jure parochiali) for forty years and more’. The prior of
Wootton Wawen protested that they had always been paid to his church until
thirty years ago, and that the people of Langley and Norton Lindsey, and of one
of the two Wolverton manors, had always gone to Wootton Wawen for baptism
and burial. The different arrangement of the last thirty years, he said, had been
agreed with one particular rector of Claverdon for the latter’s lifetime alone in
return for a yearly payment, and this had been specifically confirmed by the
bishop of Worcester. But the rector’s successor, after honouring the arrangement
for a short while, had then ignored it and begun claiming the tithes and other
revenues as Claverdon’s by ancient right.46
With the benefit of hindsight we can make sense of this ecclesiastical conun-
drum much more easily than those who were entrusted with solving it in the
thirteenth century. The clues we need are all in the records of the dispute. For
43
44
VCH, Warwickshire, III, ed. Styles, p.194, citing (n. 15) Assize Roll 948, m. 6d.
45
46
Ibid.
Page 147
Ordernr. 061297 P1
132 Steven Bassett
instance, we have the prior of Wootton’s admission that the agreement made with
a rector of Claverdon for his lifetime alone had in practice survived the latter’s
death and been carried on by his successor. This must tie in with Claverdon’s
claim to be entitled to the ecclesiastical revenues of Langley, Norton Lindsey and
Wolverton ‘by parochial right’ — that is to say, by reason of their belonging in its
parish, a claim which is fully backed up by Hugh of Hatton’s attempt to grant
Claverdon’s church and its chapels to St Florent’s in 1151x57. The arrangement
with Wootton Wawen which two successive rectors of Claverdon enjoyed was
presumably a formalized version, ratified by the bishop, of an earlier informal
arrangement.
Yet the conundrum remains. If the chapels had been built on manors in
Claverdon’s parish, why was Claverdon not automatically entitled to their
ecclesiastical revenues? What possible justification could the prior of Wootton
Wawen have had for claiming that they belonged to his church? The answer must
be that Claverdon itself had at first been subject to Wootton Wawen’s church. If
Claverdon’s own parish had developed out of a chapelry* of Wootton Wawen,
it is small wonder that when manorial chapels were later built on some of the
manors which it served on its mother-church’s behalf, a conflict of interests
should arise between Claverdon and Wootton Wawen. Viewed in this light, the
solution adopted in 1257 was eminently reasonable. Recent history, and (if such
things were ever taken into account) the best interests of the local people, dictated
that the chapels should stay subject to Claverdon; meanwhile the mother-church’s
ancient right to all the ecclesiastical revenues from its original parish would be
satisfied with a large annual payment.
Two loose threads remain. The first concerns the original status of Claverdon’s
church. The reference in Domesday Book to a priest on the manor indicates its
existence by 1086, and the evidence of Hugh of Hatton’s gift shows that by
1151x57 it was a parish church serving an area which contained six Domesday
manors. These are the attributes of a church with important origins — probably
a parochial chapel set up early on to serve roughly the eastern third of Wootton
Wawen’s original parish.
The other loose thread concerns Norton Lindsey. The suffix Lindsey derives
from the Lindsey family, who were holding the manor early in the thirteenth
century, when the existing church was built;47 it was doubtless added to distin-
guish this Norton from the many other manors of the same name. Originally
47
Gover, Mawer and Stenton, Place-Names of Warwickshire, p.216; VCH, Warwickshire, III,
ed. Styles, p.138.
Page 148
Ordernr. 061297 P1
Norton Lindsey had been called plain Norton, the ‘north tûn’, that is, ‘north farm
or manor’. Now it is not hard to work out that by no stretch of the imagination
could Norton Lindsey be said to be north of anywhere in Wootton Wawen’s
original parish. A manor with a directional name like Norton, Aston, Sutton or
Weston was so called in relation to a more important place to which it was sub-
ordinate: ‘it is our north farm, our north tûn’. To the south of Norton Lindsey is
Stratford-upon-Avon, somewhere with a church which, like Wootton Wawen’s,
was an Anglo-Saxon mother-church.48 In brief it is probable that Norton Lindsey
was at first served by Stratford’s minster, but that sooner or later it was captured
by Wootton Wawen. This occasionally happened when someone held a lot of
manors in one district, including the manor on which a mother-church stood and
also one or more manors which were subject to other nearby mother-churches. If
he gave his own mother-church the tithe from the demesne land on all his manors
in the district, and if he was never challenged by the owners of any of the other
mother-churches, he would have begun a process which might well end in the
‘capture’ of one or more manors by a church which was not their original mother.
I suggest that this explains how a manor called Norton could end up as the
easternmost part of the original parish of Wootton Wawen. To put some flesh on
these bones, Domesday Book states that in 1066 the manors of Wootton Wawen
and Norton Lindsey were both being held by a thegn, Vagn, who also had several
other manors in the same area.49 He, or someone before him who also held both
manors, arguably set this ‘capture’ in motion.
Preston Bagot and Beaudesert
Finally, I come to Preston Bagot and Beaudesert, two parishes which mirror a pair
of adjacent Domesday manors called Preston. Beaudesert’s name is Norman
French, meaning ‘beautiful wilderness’, but the history of its manorial descent
proves that it was the second of the two manors called Preston which Domesday
48
Cartularium Saxonicum, ed. Birch, II, no. 450; Sawyer, Anglo-Saxon Charters, no. 198
(Beorhtwulf’s charter of 845); Cartularium Saxonicum, ed. Birch, II, nos 533–34; Sawyer, Anglo-
Saxon Charters, no. 1278 (Bishop Wærferth’s lease of 872). The minster is also referred to in a
spurious charter of 714: Cartularium Saxonicum, ed. Birch, I, no. 130; Sawyer, Anglo-Saxon
Charters, no. 1250.
49
DB, fol. 242b; DB Warwicks., 22,9; 22,23. Vagn held at least two other manors in the area
mapped on Figs 6.1–2: DB, fol. 242b; DB Warwicks., 22,6–7.
Page 149
Ordernr. 061297 P1
134 Steven Bassett
Book records in Warwickshire.50 The proximity of these manors and their shared
name show that they must originally have formed a single land unit called ‘the
priests’ tûn’, that is, ‘farm or manor of the priests’.51 But this does not prove that
the area had ever been in Wootton Wawen’s parish: conceivably the priests of
Preston’s name could have belonged to some other mother-church. The evidence
needed to show that the entire Preston area originally belonged to Wootton
Wawen’s parish is slight and circumstantial, but it is nonetheless convincing.
It comes chiefly from the area’s location. Beaudesert and Preston Bagot are
bordered by land parochially subject to the church of Wootton Wawen (Figs
6.1–2). This occurs to an extent which can leave no reasonable doubt that they
too must originally have been subject to it. To the west lies Ullenhall, to the east
Claverdon, and to the south Wootton Wawen itself. What is more, the boundary
shared by the parishes of Wootton Wawen, Beaudesert and Preston Bagot follows
a decidedly eccentric line. There is a long, narrow tongue of land which pushes up
between the latter two parishes to within less than a kilometre of the northern
edge of Preston Bagot. It belonged to the Domesday manor of Whitley, an area
which has apparently always been in Wootton Wawen parish, even though in
terms of its natural topography Whitley is an integral part of the Preston area.52
And, so, the fact that Wootton Wawen’s church has always had parochial control
of Whitley makes it very likely that the priests of Preston’s name were those of the
minster at Wootton Wawen, and, secondly, that the now independent parishes
of Beaudesert and Preston Bagot originally belonged to its parish.
If that is correct, the two Domesday manors of Preston were presumably
formed by a subdivision of the northern part of the land owned by the minster
community of Wootton Wawen. When the minster was founded in the early
eighth century it was endowed with land assessed at twenty hides, and so it is
worth noting that Domesday Book records that in 1086 the two manors named
Preston were each assessed at five hides, the manor of Whitley was assessed at
three hides, and the manor of Wootton Wawen at seven hides.53 That is to say,
the four manors together had an assessment of exactly twenty hides in 1086. I
suggest, then, that their combined area represents what had been given to the
50
DB, fol. 240a, b; DB Warwicks., 16,18; 16, 62; Gover, Mawer and Stenton, Place-Names
of Warwickshire, p.199; VCH, Warwickshire, III, ed. Styles, p.45. Both manors were in Ferncombe
Hundred.
51
Gover, Mawer and Stenton, Place-Names of Warwickshire, p.217: Prestetone.
52
DB, fol. 242b; DB Warwicks., 22,25; VCH, Warwickshire, III, ed. Styles, pp.200–01.
53
DB, fols 240a, 240b, 242b; DB Warwicks., 16,18; 16,62; 22,9; 22,25.
Page 150
Ordernr. 061297 P1
minster at Wootton Wawen at its foundation (Fig. 6.3). We know that by 844 it
and its lands were in the hands of the cathedral church at Worcester; and in 1086
the manor of Wootton Wawen belonged to the bishop of Worcester’s private
hundred of Pathlow, even though by then it was in secular hands.54 These facts
strongly suggest that at some time in or after the ninth century the minster’s
landed endowment had been split up between the church of Worcester and the
minster’s own priests, with Worcester holding on to the land which lay closest to
the minster, and with the priests being allowed to keep the more distant part, as
a result of which the latter became known as ‘the priests’ tûn’, Preston.
The churches of Preston Bagot and Beaudesert both include a lot of
twelfth-century fabric. The former looks as if it was set up as a parochial chapel,
and Beaudesert’s has all the characteristics of a manorial one. Preston Bagot’s
church was built as a simple unicellular chapel with a minimum of architectural
or sculptural embellishment, and with much poorer materials and workmanship
than are found in any of the nearby churches of manorial origin.55 Moreover, it sat
at a considerable distance from the manor-house of Preston Bagot, located in the
most visible place available at the end of a pronounced spur of land, in the middle
of a furlong which had been cut out of one of the manor’s open fields.56 These
attributes plainly mark it out as a church that was set up, not as an act of private
devotion and self-advertisement by a lord of the manor, but as a roofed meet-
ing-place for the local population as a whole; that is to say, it too was a parochial
chapel. Almost certainly, the present church was not the first one on the site. Any
earlier one set up by the priests of Wootton Wawen was presumably timber-built,
and it could not be discovered now without extensive excavation.
The question of when Preston Bagot’s church was founded will have to stay
unanswered for the time being, but that is certainly not true of St Nicholas’s,
Beaudesert. It stands at the foot of a minor hill crowned by a ringwork-and-bailey
castle which had been built there by the Montforts by 1141.57 At first the castle
54
W. Dugdale, Monasticon Anglicanum, ed. J. Caley, H. Ellis and B. Bandinel, 6 vols in 8
(London, 1849), I, 608; DB, fol. 242b; DB Warwicks., 22,9. For the origins of Pathlow Hundred,
see Bassett, ‘Administrative Landscape’, pp.160–64.
55
S. Bassett, The Wootton Wawen Project: Interim Report No. 7 (University of Birmingham,
School of History, 1989), pp.5–12; Interim Report No. 8 (1991), pp.1–7.
56
Bassett, Interim Report No. 7, pp.12–16.
57
Regesta Regum Anglo-Normannorum, 1066–1154, 4 vols (Oxford, 1913–69), III, ed. H. A.
Cronne and R . H. C. Davis, p.220, no. 597; D. Renn, Norman Castles in Britain, 2nd edn
(London, 1973), p.104.
Page 151
Page 152
136
Ordernr. 061297 P1
Figure 6.3. The approximate extent of the land granted to the minster at Wootton Wawen at the time of its foundation
Steven Bassett
(716x737). The boundaries shown within this area indicate the likely extents of the four Domesday manors situated there.
Ordernr. 061297 P1
probably had a chapel within its earthworks, but in the 1160s a large, finely em-
bellished chapel was built on a far less restricted site just outside the castle’s outer
defences. It had its own parish by the early fourteenth century,58 but before it
became parochially independent it was probably served by Preston Bagot’s church.
That is a reasonable deduction from what we know about Beaudesert (i.e. that, as
we have seen, it was once called Preston, and that it was carved out of a larger land
unit of the same name). The parochial chapel at Preston Bagot is likely to have
served the whole of the larger area initially. This deduction is reinforced by what
is known about the glebe land* of the two churches, the post-medieval geography
of which strongly suggests that it represents a subdivision of what was originally
attached to Preston Bagot’s church alone.59
Of all the arguments made here for a manor’s original association with the
church of Wootton Wawen the one for Beaudesert, although circumstantially
sound, is the least well supported by direct evidence. And so it is important to
note that in 1535 it was reported that Beaudesert’s church paid an annual pension
of two shillings to Wootton Wawen.60 The reason for the payment is unknown,
but it provides a direct link between the two churches of a sort which normally
indicates that the payer served an area which had at first been served by the payee.
Dating the Creation of the Original Parish
That, then, is most of the evidence of which I know for the original parish of
Wootton Wawen’s church. I am fairly sure that nowhere else belonged to it,
58
Calendar of the Register of Adam de Orleton Bishop of Worcester 1327–1333, ed. R . M.
Haines, Worcestershire Historical Society, n.s., 10 (London, 1979), p.47, no. 33 (1328);
Nonarum Inquisitiones in Curia Scaccarii temp. Regis Edwardi III (London, 1807), p.448 (1341).
It may already have been parochially independent by 1283, when the first of several references
occurs to a ‘rector of the church of Henley’, which is almost certainly St Nicholas’s, Beaudesert:
A Descriptive Catalogue of Ancient Deeds, 6 vols (London, 1890–1915), III, 82, no. A 4572;
Calendar of Patent Rolls, 1292–1301 (London, 1895), p.282 (1297); The Register of the Diocese
of Worcester during the Vacancy of the See, usually called ‘Registrum Sede Vacante’, ed. J. W. Willis
Bund (Oxford, 1897), p.65 (1302).
59
Ecclesiastical Terriers of Warwickshire, vol. II, transcribed and ed. D. M. Barratt, Dugdale
Society, 27 (1971), pp.39–42; Warwicks. Record Office, CR 569/23 and CR 569/192 (tithe
apportionment and map for, respectively, Beaudesert and Preston Bagot); Bassett, Interim Report
No. 8, pp.6–8.
60
Valor Ecclesiasticus temp. Henr. VIII auctoritate regia institutus, 6 vols (London, 1810–34),
III, 91.
Page 153
Ordernr. 061297 P1
138 Steven Bassett
because I have done similar work in all the adjacent areas and have found that the
manors and ecclesiastical parishes there belonged to the original territories of
other mother-churches.61 So, finally, a few words are needed about the light that
this case study throws on the fiercely debated issue of the date at which minsters
of early foundation like Wootton Wawen acquired a parish. Was it indeed not
until the late Anglo-Saxon period, at the hands of one of the early rulers of a
united kingdom of England, such as Athelstan or Edgar? Or did these minsters
have parishes at a much earlier date?
Not one of the subordinate churches discussed above can be shown to have
existed before the eleventh century, and most of them may well not have done so.
Similarly, in almost every instance the evidence produced here of a now indepen-
dent parish’s former subjection to Wootton Wawen can be comfortably inter-
preted in terms of a parochial system which had been newly set up in the late
Anglo-Saxon period. This is exactly what I have found in the many other studies
which I have made of the original extents of mother-church parishes in the west
midlands — that in almost every case there is no evidence of mother-daughter
relations which cannot have arisen in or after the tenth century. This is hardly
surprising given the late date of most of the evidence that has to be used; but there
is no room here for special pleading. If the debate about the origins of the par-
ochial system is to be moved on — that is to say, if we are to make significant
progress in discovering if the mother-church parishes whose fission produced the
parochial geography of late medieval England were ones newly created in or after
the tenth century, or if they were of much greater antiquity — then clear signs
must be sought of mother-daughter relationships which can have arisen only
before the tenth century.62
Evidence to this effect has already been found in several parts of the west
midland region.63 There seems to be similar evidence for the Wootton Wawen
area too. To examine it we need to make a very brief return to Oldberrow, the one
part of this whole area which in 1086 was in Worcestershire, not Warwickshire
61
Bassett, ‘Administrative Landscape’, pp.160–64.
62
This presupposes that such relationships were never imposed on pre-existing churches
which, prior to the granting of matronal rights and responsibilities to one church within a given
district, had been in no way dependent on the latter. I admit to holding this presupposition,
believing there to be substantive, though admittedly only circumstantial, arguments in support
of my position, which are, however, too lengthy to be set out here.
63
For example, Bassett, Origins of the Parishes of the Deerhurst Area, pp.6–7 (Bishops Cleeve),
8–9, 21–24 (Beckford), and 20; Bassett, ‘Anglo-Saxon Birmingham’, pp.10–12 (Aston).
Page 154
Ordernr. 061297 P1
(because it was held then by Evesham Abbey). Moreover, it lay in Evesham’s small
private hundred, Fishborough Hundred, unlike many of the abbey’s other
manors, including less distant ones.64 This fact and its eccentric membership of
Worcestershire65 firmly indicate that Oldberrow was already in Evesham’s hands
at the time of the tenth-century reorganization of secular administration in the
former kingdom of Mercia which was undertaken by its new West Saxon rulers.
(I argued earlier from Beorhtwulf’s alleged privilege that it had probably been
Evesham’s since at least the mid-ninth century.)
Yet it is certain that Oldberrow was parochially subject to Wootton Wawen’s
church, and that all Evesham’s efforts to undermine this relationship in and after
the twelfth century were in vain.66 This might mean no more than that in the
tenth century some bright young West Saxon policy-maker devised a genuinely
community-friendly plan for delivering pastoral care to everybody in England —
someone who, swimming against the strong tide of monastic vested interests, was
determined to ensure that every one of his new mother-church parishes should be
geographically coherent, no matter how often it meant treading on the toes of a
major church like Evesham. But if so, it is amazing that his grand plan should have
been thwarted so often elsewhere in England — as it plainly was, if that is how
things really were — and yet should have proved so robust at Oldberrow. A more
plausible explanation is that Wootton Wawen’s church had already got parochial
control of Oldberrow before it came into Evesham’s hands, which (as we have
seen) was probably not later than the 840s. This explanation certainly puts a lot
of faith in the reliability of an entry in an Evesham cartulary, which is not nor-
mally a wise move. But it is very hard to believe that the abbey’s monks would have
seen any value in spuriously naming the obscure Beorhtwulf as a benefactor. And
so I suggest that what the charter says about him is true, and that we may take it
that Oldberrow was already subject to the church of Wootton Wawen by, at the
latest, the end of his reign in 852. If it was, this looks like another example of a
mother-church which very probably had a parish before the tenth century.
64
DB, fol. 175b; Domesday Book, vol. XVI: Worcestershire, ed. F. Thorn and C. Thorn
(Chichester, 1982), 10,4. Lying between Evesham and Oldberrow are four Evesham Abbey
manors which were not in Fishborough Hundred: ibid., 10,13–16.
65
Evesham Abbey’s manors at Abbot’s Salford and Weethley lay between it and Oldberrow
but were situated in Warwickshire (DB, fol. 239a; DB Warwicks., 11,3; 11,5), even though they
immediately adjoined Worcestershire.
66
Above, pp.125–27.
Page 155
Ordernr. 061297 P1
140 Steven Bassett
Conclusion
That is by no means a watertight conclusion, and plainly it will not put an end to the
fierce debate about when Anglo-Saxon minsters gained parishes. However, it was
never my intention to do so here. Instead, I have sought to prove the reality of one
particular late Anglo-Saxon mother-church’s parish and to chronicle its slow
fission.67 While it is still in our sights, one other aspect of this area is worth con-
sidering. Figure 6.4 shows it superimposed on the local drainage pattern. It is im-
mediately clear that the district which Wootton Wawen’s church initially served
comprised the middle section of the drainage basin of the River Alne, with its north-
western boundary mainly coinciding with the watershed between the Alne itself
and the River Arrow, and its south-eastern boundary (if Norton Lindsey is excluded)
coinciding with the one between the Alne’s easternmost tributary and the River
Avon. The other two boundaries lie at right-angles to the predominant trend of
the Alne and its main tributaries. The area thus enclosed would have contained
a cross-section of land-use types, offering the resources needed for a mixed economy
based on arable cultivation, stock rearing and the exploitation of woodland.
As an undivided territory it is reminiscent of the early Anglo-Saxon ‘river
estates’ which have been identified elsewhere in lowland Britain.68 Within it lay
the twenty hides of land which Æthelbald ‘booked’ to the minster at Wootton in
716x37, describing it as ‘all this area within the boundaries set up by its ancient
possessors, with fields and woods and meadows and everything which can be pro-
duced from it for the necessities of life’.69 This endowment was ‘in fact an area
67
That parish almost certainly had neither the full extent of the area depicted in Fig. 6.2 nor
internal secular boundaries which exactly followed the lines of the parish boundaries shown there.
Norton Lindsey was probably not an original component of it (above, pp.132–33); similarly,
Morton Bagot (which lies in the Arrow valley, not that of the Alne) cannot be ruled out as
another possible example of ‘capture’. Some of the internal boundaries mapped in the early
nineteenth century may represent the end product of a process of internal colonization which was
still incomplete in the late Anglo-Saxon period; and blocks of land — often only single fields —
are likely to have been transferred from one parish to another for tenurial or other reasons, as
happened to the western half of the Domesday manor of Kington (above, note 40), for example.
68
For example, W. G. Hoskins and H. P. R . Finberg, Devonshire Studies (London, 1952),
pp.302–05; T. R . Slater, ‘More on the Wolds’, Journal of Historical Geography, 5 (1979), 213–18;
A. Everitt, Continuity and Colonization: The Evolution of Kentish Settlement (Leicester, 1986),
ch. 4.
69
‘Omnem itaque hunc agrum intra terminos ab antiquis possessoribus constitutos, cum
campis et saltibus et pratis necnon et uniuersis qui ex eo provenire possint necessariis redituum
usibus’: Cartularium Saxonicum, ed. Birch, I, no. 157; Sawyer, Anglo-Saxon Charters, no. 94.
Page 156
Page 157
Ordernr. 061297 P1
Figure 6.4. The probable maximum extent of Wootton Wawen’s parish

141
shown in relation to natural drainage.

Ordernr. 061297 P1
142 Steven Bassett
which lies in the territory anciently called Stoppingas’; and the natural topography
of the district within which the minster and its land were situated has a scooped-
out appearance which invites identification as the putative ‘bucket-shaped hollow’
(*stoppa) of that name.70 That is to say, the obvious topographical coherence of
the area which, arguably, Wootton Wawen’s church was serving by the mid-ninth
century suggests that it became a formal territory, not by the aggregating of a
cluster of separate manors for ecclesiastical purposes, but through its being a
primary unit of early Anglo-Saxon settlement and land use.
The case study at the heart of this essay has produced a solid, largely secure
exemplification of the model for the origins of the parochial system which was set
out at the start, and there is no good reason to regard Wootton Wawen as excep-
tional in any important respect, not even in terms of the quantity or quality of the
surviving evidence used in this study. In the west midlands, as elsewhere, a great
many of the late Anglo-Saxon mother-churches can be reliably identified; and,
where appropriate evidence exists, their original parishes can be wholly or partly
mapped and their eventual disintegration tracked.71 This gives the study of local
communities an essential topographical dimension which, amplified by further
historical research and, where appropriate, by fieldwork and other archaeological
activities, can greatly enhance our understanding of how they functioned, inter-
acted and evolved. This is likely to be as achievable in respect of specific secular
communities — members of an English manor, say, or hundred — as of an
ecclesiastical one. Mapping medieval land-units is an excellent way of combining
and thereby enhancing what we know of the ecclesiastical, socio-economic, and
political/administrative organization of the inhabitants of any given area.72
70
‘Est autem idem ager qui traditur in regione quae antiquitus nominatur STOPPINGAS.’
I am grateful to Margaret Gelling for discussing this name’s etymology and historical significance
with me, and for her endorsement of my suggestion that ‘bucket-shaped hollow’ is topographically
appropriate for the area in question (pers. comm., 29 July 1998).
71
See note 11 for published examples.
72
I am most grateful to Wendy Davies, Julio Escalona, Duncan Probert and Orri Vésteinsson
for their comments on an earlier draft of this chapter, and to Donald Graham for advice on the
extent of the late medieval manor of Offord.
Page 158
Ordernr. 061297 P1
M APPING S CALE C HANGE: H IERARCHIZATION

AND F ISSION IN C ASTILIAN R URAL C OMMUNITIES
DURING THE T ENTH AND E LEVENTH C ENTURIES
Julio Escalona
R
ural communities, their social life and relationships are prominent in Castile
north of the River Duero, where surviving written sources from the period
before 1100 begin earlier and are more abundant than those from south of
the river. This material makes a suitable point from which to observe medieval com-
munities, and one that bears interesting comparisons with other areas in the Iberian
northern fringes, such as Catalonia or Galicia, and elsewhere in Europe, as other
essays in this volume make clear. However, the peculiarities of medieval Iberia’s histo-
riography loom high over any attempt to compare Castilian communities with their
European counterparts. This is true of many other areas — witness, again, most
of the essays in this book — but a strong caveat is nevertheless necessary for early
medieval Castile, because of the present crisis of a whole historiographical model
based upon the repopulation of deserted — or very thinly populated — areas. Since
the 1990s new views have emerged, but they are only just beginning to be credited
as anything more than marginal in the most general state-of-the-art reviews, and are
clearly under-represented in the available Spanish, let alone English, literature. It
seems appropriate, therefore, to start with a brief presentation of recent views of Cas-
tilian space and communities to help contextualize the material I will discuss later on.
The Historical/Historiographical Setting
The Historical Process
Following the Visigothic collapse in 711, effective Arab control came, by 770, to
be restricted to the lands south of the Spanish Central mountains and the
Page 159
Ordernr. 061297 P1
144 Julio Escalona
Figure 7.1. North-west Iberia in the tenth-century:

the Asturian kingdom and the Castilian county.
Pyrenean piedemont. Of the northernmost reaches, the east soon became involved
in Carolingian expansion. Much less so the west, where a kingship of obscure
origin developed in the eighth century in the Asturian uplands, managed to gain
a certain stability during the first half of the ninth century, and in the late ninth
to early tenth century expanded over the northern half of the Duero basin to
become the largest Christian polity in Iberia. Meanwhile the territories between
the Duero and the Cantabrian mountains lay beyond the control either of the
Cordoban emirs* or the kings of Oviedo (Fig. 7.1). Late ninth-century chron-
iclers seeking to legitimize the Asturian take over argued that these lands had
remained depopulated after 711. This idea was to influence interpretations of
medieval Spain up to the mid-twentieth century.1
The eastern side of the kingdom of Oviedo-León was the county of Castile.
Originally a small territory just south of the Cantabrian mountains, Castile
retained its name through a process of expansion southwards that repeated, a
generation later, that which had taken place on the western side of the plateau. By
930, Castile comprised the lands from the Cantabrian sea to the River Duero,
between the River Pisuerga and the Iberian Mountains that divide the plateau
1
J. Escalona, ‘Family Memories: Inventing Alfonso I of Asturias’, in Building Legitimacy:
Political Discourses and Forms of Legitimation in Medieval Societies, ed. I. Alfonso, H. Kennedy
and J. Escalona (Leiden, 2004), pp.223–62.
Page 160
MAPPING SCALE CHANGE 145
from the central Ebro valley (Fig. 7.1). This is the territory to which I will refer as
Castile in this text. From 931 to 1027, it was ruled by the Castilian comital
dynasty under the sway of the Leonese kings, and in certain periods it enjoyed a
remarkable degree of autonomy.
Historiography: From Vacuum to Continuity
Faced with the difficulty of accounting for the Duero basin’s ill-documented
history between 711 and the Asturian-Leonese takeover, modern historians have
relied on explanations largely cast in the mould of the ninth-century royal
chronicles. This was taken to its most sophisticated form by Sánchez-Albornoz,
establishing a dominant orthodoxy for most of the twentieth century.2 According
to him, the Duero plateau had been deliberately devastated in the eighth century
by King Alfonso I to create a defensive no-man’s-land between his kingdom and
Al-Andalus, and was gradually occupied by the Asturian repopulation in the fol-
lowing centuries. On this allegedly empirical basis Sánchez-Albornoz defined the
Asturian kingship’s ultimate historical mission: to preserve the Visigothic institu-
tional inheritance and pass it on to a social body and territory whose links with
the Roman and late antique past had been abruptly cut out and replaced with a
new society of freemen; the influence of Turner’s American Frontier on Sánchez-
Albornoz has long been recognized.3 The thin archaeological evidence and the fact
that the region’s post-Visigothic texts start with its new Asturian masters —
mainly great monasteries — helped the depopulationist theory become almost
untouchable. All fresh evidence thereafter was used in circular arguments, inter-
preted in the light of this uncontested orthodoxy, and then used to reinforce it.4
Perhaps surprisingly, given the great theoretical importance attached to it, settle-
ment itself was little investigated until the 1960s. Historians were more concerned
with legal history and the Visigothic origins of the Asturian institutions; post-
Visigothic medieval archaeology was practically non-existent. Then in the 1960s
a new breed of studies tried to map the milestones of Christian ‘repopulation’
2
C. Sánchez-Albornoz, Despoblación y Repoblación del Valle del Duero (Buenos Aires, 1966).
3
On Sánchez-Albornoz’s legacy, see R . Pastor, C. Estepa, J. A. Garica de Cortázar, J. L.
Abellán and J. L. Martín, Sánchez Albornoz a debate (Valladolid, 1993).
4
F. Reyes and M. L. Menéndez Robles, ‘Aspectos ideológicos en la despoblación del Valle del
Duero’, in Historiografía de la Arqueología y de la Historia Antigua en España (siglos XVIII– XX )
(Madrid, 1991), pp.203–07.
Page 161
Ordernr. 061297 P1
146 Julio Escalona
Photo 7.1. The Liébana valley, Cantabria.
southwards,5 mostly dots on maps of the struggles between Christian and Muslim.
It was only in the 1970s and 1980s that settlement patterns became the subject of
high-quality research, mainly based on written and place-name evidence.6
By 1990, most scholars would agree upon a threefold scheme of Castilian-
Leonese settlement: 1) In the northern mountainous lands — neither controlled
by the Arabs nor depopulated by the Asturian kings — the dominant pattern was
that of ‘valley communities’. Those were substantial upland territories occupied
by scatters of farm-type dispersed settlements, with churches as territorial foci,
and a network of social ties across the valley, mostly based upon common manage-
ment of, and access to, uncultivated land and other resources (pasture, wood,
5
S. de Moxó, Repoblación y Sociedad en la España cristiana medieval (Madrid, 1979).
6
Especially J. A. García de Cortázar, ed., Organización social del espacio en la España Medieval:
La Corona de Castilla en los siglos VIII al XV (Barcelona, 1985); J. A. García de Cortázar,
‘Organización social del espacio burgalés en la Alta Edad Media’, in II Jornadas Burgalesas de
Historia: Burgos en la Alta Edad Media (Burgos, 1991), pp.19–74; P. Martínez Sopena, La Tierra
de Campos Occidental: Poblamiento, poder y comunidad del siglo X al XIII (Valladolid, 1985).
Page 162
Ordernr. 061297 P1
water and so on). During the central Middle Ages, individual villages — nucleated
or not — were defined within most of those valley units, but the larger territories
stayed the main framework for much of social life.7 2) On the plateau north of the
Duero — the limit of tenth-century Christian expansion — the settlement
pattern was dominated by a spread of ‘village communities’ of local scale and peer
rank, with neither hierarchies nor any higher articulation amongst them. In the
1980s these were mostly seen as the result of spontaneous peasant colonization of
deserted spaces. Yet it is the process by which local communities were controlled
thereafter by a landed, mainly ecclesiastical aristocracy that is most clearly evi-
denced in tenth- to eleventh-century charters. Subsequently, in the second half of
the eleventh century, the articulation of the parish* network confirmed those
village communities as the dominant element in the central Castilian lands.
3) Finally, in the lands between the river Duero and the Central mountains,
which were added to the Leonese kingdom in the second half of the eleventh
century, the pattern was based on central places of varying rank — from cities
with an episcopal see to very small towns — each with a hinterland containing a
number of dependent settlements (aldeas).
As a result, north and south territoriality was conceived in two levels: local
settlements embedded in larger territories, whether ‘valley communities’ or town-
ships (with central places in the south, without them in the north), but in central
Castile lay a ‘flat’ mass of peer local settlements, which certainly has a bearing on
the degree of social complexity that can be postulated for it. Moreover, links with
the late antique regional past remained hazy: in the north, there was continuity,
but from a number of allegedly ill-Romanized, largely indigenous societies (‘tribal
structures’ was a fashion of the time). In both the middle and the south, due to a
more or less prolonged period of desertion, traces of earlier settlement were either
lost or negligible, and overlaid with the new ones. Finally, the bulk of peer rural
settlements in the plateau north of the Duero were seen as resulting from
spontaneous peasant colonization, followed by the establishing of a network of
administrative districts by higher political powers, and the emergence of towns
and large seigneurial dominions. South of the river, and much later, the kings were
7
C. Díez Herrera, La formación de la sociedad feudal en Cantabria (Santander, 1990);
I. Martín Viso, ‘Poblamiento y sociedad en la transición al feudalismo en Castilla: Castros y aldeas
en la Lora burgalesa’, Studia Historica: Historia Medieval, 3 (1995), 3–45; J. M. Imízcoz,
‘Comunidad de valle y feudalismo en el norte de la península: Algunas preguntas desde el Valle
de Baztán’, in Señorío y feudalismo en la Península Ibérica (siglos XII– XIX ), ed. E. Sarasa Sánchez
and E. Serrano Martín, 4 vols (Zaragoza, 1993), III, 69–86.
Page 163
Ordernr. 061297 P1
148 Julio Escalona
supposed to have created from scratch an articulate network of townships and

their hinterlands to which the basic rural settlements were ascribed.8
New Perspectives
Although still favoured by many, the threefold pattern of the 1980s nowadays
arouses much less vigorous enthusiasm. Since the mid-1990s, new research has
called for the consideration of long-term dynamic continuity, as well as greater so-
cial complexity.9 The new views owe much to the growing wave of criticism, from
the late 1970s, of the textual evidence upon which the depopulation-repopulation
model was based and the dangers of taking its discourse at face value.10 Closer
investigation of the earliest recorded territorial districts (the so-called alfoces*) has
also played an important role.11 Much less so archaeology — admittedly the main
potential source of fresh evidence — through a lack of tightly-focussed large-scale
excavations, although it should take the lead in the near future.12
Thin as it is, the available evidence can be used to produce an image of early
medieval settlement patterns in central Castile that is rather different from the
standard model of the 1980s. My aim will be to suggest that early Castilian local
communities were not just scattered peer settlements, but belonged to a more
complex territorial pattern, with settlements of different rank and a network of
supralocal territories which largely influenced the formation of the earliest-recorded
administrative districts. Moreover, in the period I will consider (the tenth and
8
Compare Astill, this volume.
9
J. Escalona, Sociedad y territorio en la Alta Edad Media Castellana: La formación del alfoz de
Lara, British Archaeological Reports, International Series, 1079 (Oxford, 2002); I. Martín Viso,
Poblamiento y estructuras sociales en el norte de la Península Ibérica: siglos VI– XIII (Salamanca,
2000); I. Martín Viso, Fragmentos del Leviatán: La articulación política del espacio zamorano en
la alta edad media (Zamora, 2002), among others.
10
A. Barbero and M. Vigil, La formación del feudalismo en la Península Ibérica (Barcelona,
1978).
11
C. Estepa Díez, ‘El alfoz castellano en los siglos IX al XII’, En la España Medieval, IV:
Estudios Dedicados al Profesor D. Ángel Ferrari Núñez, 2 vols (Madrid, 1984), II, 305–42;
I. Álvarez Borge, Monarquía feudal y organización territorial: Alfoces y merindades en Castilla
(siglos X – XIV ) (Madrid, 1993).
12
The dominant picture of the 1980s, considerably more complex and nuanced than its
predecessors, contained very little input from medieval archaeology, which, despite its rapid
development early in the decade, hardly entered historians’ discourse — nor academic posts.
Page 164
Ordernr. 061297 P1
eleventh centuries) the whole pattern was subject to dramatic changes, of which the
most remarkable was the fission of village territories, fully comparable to Steven
Bassett’s discussion, in this volume, of the evidence of Anglo-Saxon parishes.
I will revisit here some cases I have discussed elsewhere together with other
freshly assessed material, combining charters with place-names, topography, archi-
tecture and, when available, archaeological survey data.13
Village Fission: A Topographical Exploration
The Castilian medieval parish network crystallized in the late eleventh century,
and most of it has survived up to the present. In south-eastern Castile one parish
normally corresponds to one village, many of which can be traced back to the
tenth century. At first sight this would suggest that medieval parishes drew
directly on a settlement pattern in which village territories were the principal
building blocks. However, closer inspection reveals that the late eleventh-century
pattern results from the fission of larger, supralocal units* into their component
settlements. This process was already on its way when our first charters were
written down,14 so, instead of revealing a starting point, what the textual evidence
shows is in fact glimpses of ongoing changes. Moreover, fission worked unevenly,
at different paces and in varying intensity in different areas. The territories of
Ausín and Orbaneja serve to illustrate this variability.
Village Fission in the District of Ausín
In the tenth and eleventh centuries, Ausín was a small alfoz some 15 km south-
east of Burgos, formed by a stretch of river with other secondary valleys (Fig. 7.2).
Charter evidence is available from the 940s. This must be combined with the
study of place-names, topography and the extant architecture. Despite the poten-
tially rich archaeological remains, no excavation has been undertaken hitherto,
only non-systematic field surveys, focussed on the pre-Roman and Roman
13
Vésteinsson’s and Bassett’s opening discussions in this volume are generally pertinent here,
especially regarding handling of place-name evidence.
14
Charter writing was itself a chief factor of the process. See J. Escalona, ‘De “señores y
campesinos” a “poderes feudales y comunidades”: Elementos para definir la articulación entre
territorio y clases sociales en la Alta Edad Media castellana’, in Comunidades locales y poderes
feudales en la Edad Media, coord. I. Álvarez Borge (Logroño, 2001), pp.117–55.
Page 165
Ordernr. 061297 P1
150 Julio Escalona
Figure 7.2. The district of Ausín in the tenth to eleventh centuries.

(Arrows indicate intrusions of neighbouring districts into Ausín’s.
Village-fission cases are grouped by dashed lines.)
periods.15 Around the central place, Ausín, there lay four smaller territories, largely
defined by the drainage network: Modúbar, Hontoria, Revilla and Cubillo. By the
15
For a discussion of the evidence for medieval Ausín, see Escalona, Sociedad y territorio,
pp.94–110.
Page 166
Ordernr. 061297 P1
end of the eleventh century, all four had split into two or more village territories.
This process can be investigated through place-names and landholding patterns.
Place-name evidence is quite consistent. In all four units, the individual
settlements had twofold names. The first element was the same for each area, and
villages were singled out by adding a second term connected by the preposition
‘de’.16 Thus, in the small valley of Modúbar, north of Ausín, there lay Modúbar
de San Cebrián, Modúbar de Zahalanes, Modúbar de la Cuesta and Modúbar de
la Emparedada. Likewise, in the south-western side of the district, the settlements
along the watercourse were Cubillo del Campo, Cubillo del César, and Cubillejo
de Lara. Just east of Ausín the same phenomenon reappears: Revilla del Campo
and Revilla de la Fuente. And, again, south-west of Ausín, there were Hontoria
de Suso and Hontoria de Yuso (Fig. 7.2).
Charter references suggest that the larger areas were originally the basic terri-
torial framework within which village territories defined themselves. In the tenth
century most charter references to the Modúbar area merely cite ‘in Motua’. A
church of San Cebrián is mentioned in 944, 963 and 964, but it is not until 1060
that we first see the village called ‘Mutuba de Sancti Cipriani’. This does not mean
that village territories did not exist before; a more eloquent text of 978 reads: ‘in
loco que dicitur Motua in villa de Zafalanes territorio fundate concedimus terras,
vineas, mulinos’, meaning that we are giving arable, vineyards and mills in the
deserted village of Modúbar de Zahalanes, near Modúbar de San Cebrián. It
rather seems that there were two territorial layers, and more often than not the
supralocal one was seen as fit for mention in charters. The process of individuali-
zation, though, was uneven. Revilla del Campo is recorded in 1011 with its pres-
ent name. So are Modúbar de San Cebrián in 1060, and Modúbar de Zahalanes
in 1077. By contrast, in the Cubillo valley, villages are first singled out by a
personal name (‘Cubillo de Sesnando’ and ‘Cubillo de Don Cipriano’) before they
reach their final form, sometime before the late twelfth century.17
The analysis of land transfers allows for a more dynamic approach. Landhold-
ings in tenth-century Castile were typically fragmented. Large, unitary estates
were extremely rare; more normally an individual owned a set of small, dispersed
units. If those pieces were scattered across a supralocal territory, it is conceivable
that individual estates would extend over village boundaries. This is shown quite
16
Compare Bassett, this volume.
17
Burgos, no. 34 (1077): ‘in Cupiello de Sesnando, illa uestra portione; et in Cupiello Domini
Cipriani, illa uestra portione’. In 1053 another text cites ‘Cupiello de Nofar’ which may or may
not be one of the aforesaid (Burgos, no. 17).
Page 167
Ordernr. 061297 P1
152 Julio Escalona
consistently in land transactions. While the aforementioned 978 charter mentions

land in the village territory of Zahalanes within the locus called Modúbar, the
properties exchanged are more often located by mere reference to the larger territory
(1011: ‘in Mutua nostram porcionem’, and likewise in 1029, 1039 and 1044; also
Hontoria in 1029).18 Alternatively, the existence of more than one homonymous
settlement begins to appear in texts from 1011, sometimes with formulae* like ‘in
Fonteaurea nostram porcionem et in alia Fonteaurea nostram porcionem’ (our
share in Fonteaurea and our share in the other Fonteaurea) (1029) or ‘in Cupiello
et in alio Cupiello’ (in Cupiello and in the other Cupiello); others with generic
mentions in the plural: ‘in duos Cupiellos’ (in the two Cupiellos), ‘in ambas Fonte
Aureas’ (in both Fonte Aureas), ‘in ambos Cupiellos’ (in both Cupiellos).19 From
the mid-eleventh century, it is increasingly common to see settlements mentioned
individually (1060: ‘in Mutuba de Sancti Cipriani, meas kasas’), but references in
the plural persist (1194: ‘quantum habemus in Ribiella de Campo et in ambas
Fontorias’; 1225: ‘en las Fontorias [. . .] en las Muduvas’; 1246: ‘quanto nos per-
teneçe en ambas las Fontorias’) suggesting that landholdings formed a continuum
throughout the old supralocal units, regardless of village boundaries.20
The topography confirms this impression. In all four units, the basic layout of
watersheds and drainage provide self-evident boundaries that correspond to those
of the supralocal territories. The dominating axial streams favour a land-use
sequence in bands (wet riversides, arable in the valley flatlands, pasture and
vineyards in the slopes) flowing continuously along the valley. Village boundaries
cut across this landscape, defining valley segments whose territoriality is much less
obvious than that of the larger units. Moreover, in at least two cases (Modúbar
and Cubillo) the supralocal unit and its main watercourse were homonymous.
18
Cartulario de San Juan de la Peña, ed. A. Ubieto Arteta (Valencia, 1962), no. 49 (1029):
‘in Fonteaurea de Agusyn meam portionem’. This being a substantial bequest of property over a
very wide area, the added term ‘de Ausín’ is meant here to differentiate Ausín’s Hontoria area
from other Hontorias in the southern Castilian region (the name is not that unusual).
19
Compare the cases of plural place-names in S. R . Bassett, ‘Continuity and Fission in the
Anglo-Saxon Landscape: The Origins of the Rodings (Essex)’, Landscape Archaeology, 19 (1997),
25–42, and Vésteinsson this volume.
20
As late as 1352 both Hontorias are recorded as separate villages, but they have exactly the
same seigneurial structure, and a rather complex one, with many noble persons sharing rights in
them. The same pattern seems to have applied to the Revilla area, but in this case one of the
settlements was transferred to ecclesiastical lordship, and this originated diverging seigneurial
models for both of them (Libro Becerro de las Behetrías, ed. G. Martínez Díez (León, 1981), XI,
61, 62, 107, 108).
Page 168
Ordernr. 061297 P1
Photo 7.2. Ausín.
The paucity of archaeological evidence makes it hard to explore the chrono-

logical depth of this model for all units, except Hontoria. This corresponds to the
gentle little valley of the River Saelices. Here valley and territory do not have the
same name. Hontoria derives from the Latin Fons Aurea, admittedly denoting a
water-related sacred place, while Saelices comes from San Félix (see below). As we
have seen, some charters sometimes merely mention ‘Hontoria’ and others ‘the
Hontorias’ or ‘both Hontorias’, but finally they come to distinguish two settle-
ments: Hontoria de Suso and Hontoria de Yuso (meaning Upper and Lower
Hontoria, respectively), although joint mentions of both villages keep appearing
eventually.
This picture can be combined with other evidence, such as religious architec-
ture. In Hontoria de Yuso a charter of 105621 mentions a lost church, which is
consistent with its typically early dedication to Santa Eugenia. Further up the
21
Cartulario de San Pedro de Arlanza (antiguo monasterio benedictino), ed. L. Serrano
(Madrid, 1925), no. 56 (1056). The early date surely makes the case for a building in the pre-
Romanesque tradition.
Page 169
Ordernr. 061297 P1
154 Julio Escalona
valley Hontoria de la Cantera’s church is placed on a mound overlooking the

modern village and the whole valley. Again, it bears a typically early medieval
dedication (St Michael), but the church was rebuilt in the Renaissance, leaving no
trace of earlier structures.
However, there is yet a third church in the area, on a discrete elevation amidst
the valley lowlands between both Hontorias, closer to the upper settlement. Only
a mid- or late twelfth-century Romanesque apse remains of a building that once
lay over a substantial Roman site. The nave’s walls were torn down in the first half
of the twentieth century to retrieve an ensemble of first-century funerary slabs of
great size, after which the nave virtually disappeared.22 While the slabs bear
witness to a substantial early Roman settlement, pottery finds from all around the
area clearly indicate occupation in the late empire. The evidence is not enough to
reveal what type of site there was there — a villa*-building and a nucleated village
are equally plausible — or to date its final occupation more accurately. But in the
absence of a more defined archaeological context, we can sort the available
evidence into a rough stratigraphic order.
The Hontoria area seems to have contained a substantial settlement in Roman
times, then developed into a single territory. The name Fons Aurea may well date
back to that period, and eventually have come to designate the whole valley. In the
post-Roman period this site, whatever its character, remained the main local focus
and was eventually marked with a church dedicated to St Felix, which gave its
name to the River Saelices but not to the individual settlements ([ecclesia] Sancti
Felicis > San Felices > Saelices).23 At an unspecified period, secondary settlements
developed up and down the valley, both with churches with potentially early
medieval dedications. Moreover, San Miguel’s topographical setting and the
reference to Santa Eugenia in a charter of 1056 both suggest pre-Romanesque
buildings, although we cannot be precise about how old they were in the eleventh
century, when the fission process begins to be expressed in our documents.
The Hontoria case sheds some light on the rest of the Ausín area. Supralocal
units seem to have preceded individual villages and were more relevant than the
latter to the pattern of territorial organization. Fons Aurea may well have been the
name of the Roman site (and perhaps of the whole valley); the name lived on long
enough to be passed on to the medieval villages; the axial stream may well have
22
J. A. Abásolo Álvarez, Epigrafía romana de la región de Lara de los Infantes (Burgos, no.
1974), pp.35–39.
23
On this particular place-name, see R . Menéndez Pidal, En torno a la lengua vasca (Buenos
Aires, 1962), pp.67–71, but see A. Porlan, Los nombres de Europa (Madrid, 1998), pp.577–88.
Page 170
Ordernr. 061297 P1
shared that name, but this was later replaced by that of the church which gave
continuity to the erstwhile dominant focus in the valley. The Roman site in
Hontoria de la Cantera/San Félix, together with the less significant Roman finds
in Revilla del Campo24 could also indicate that sometimes — but not always —
one village may have been central to the area before fission took place. (There are
hints that Modúbar de San Cebrián dominated the whole Modúbar unit, but the
Cubillo valley looks like an undifferentiated linear arrangement.)
However that may be, in the tenth and eleventh centuries the supralocal units in
the Ausín area were giving way to village territories. As in some of the Icelandic farms
discussed by Orri Vésteinsson and Birna Lárusdóttir in their respective essays, traces
of the larger unit survive in the form of a place-name that is shared among several
settlements. These traces can also be found in the landscape, in land use and in
landholding and seigneurial patterns. But people lived in villages, which ultimately
became the basic units for secular and ecclesiastical administration. As far as gov-
ernment was concerned, people belonged to two territorial layers: the alfoz (district)
of Ausín and their respective villages. The mechanisms of social and political inte-
gration beyond these levels remain a matter for future research. As for the smaller
territories, perhaps the supralocal units faded out very quickly and silently, or
perhaps we just lack the evidence to assess their persistence. In this respect, the
case of Orbaneja provides an interesting counterpart to Ausín.
Orbaneja: The Villas and the Valley
From 942, the cartulary* of San Pedro de Cardeña — an important monastery

about 9 km south-east of Burgos — shows the house gaining property in a nearby
area called Orbaneja. This corresponds to the upper half of the valley of the River
Picos up to the ridge known as Sierra de Atapuerca (Fig. 7.3). Just as Modúbar or
Hontoria were parts of the district of Ausín, Orbaneja belonged to the district of
Burgos, which, being a much more important unit, was much more likely to
embed smaller territories within its own (see below).
Place-names here clearly do not fit the Ausín model. The name Orbanelia >
Orbaneja could just as well derive from a personal name *Urbanus25 as from an
24
Escalona, Sociedad y territorio, pp.96–97.
25
Gregory of Tours (Historiae, X.31) mentions a church in Orbaniacum > Orbigny, the kind
of -iacum place-name that Halsall discusses in this volume.
Page 171
Ordernr. 061297 P1
156 Julio Escalona
Figure 7.3. The Orbaneja valley in the tenth to eleventh centuries.
obscure *Orbana, perhaps of pre-Roman origin.26 Whatever its origin, the

area’s name was not shared by its main watercourse, the River Picos. Neither did
the individual settlements within it take the area’s name. In 1073 they were called
Orbaneja de Picos, Villa de Valle, Villa de Plano, Villa de Domna Eilo and
Cardeñuela. The term villa is standard medieval Castilian for ‘rural settlement’,
so this aspect is that which is most emphasized in Orbaneja. ‘Cardeñuela’ must be
a development that belongs later than the beginnings of Cardeña abbey’s influ-
ence in the area. As for Orbaneja de Picos, it fits a less usual pattern: area name +
de + river name. This obviously makes sense only if the area and the river do not
have the same name and if this Orbaneja is — and, as it happens, it is — the only
settlement in the area with this name. The system could be argued to be distin-
guishing itself from other places called Orbaneja elsewhere in the Burgos hin-
terland so, again, it could be a relatively late feature, largely imposed from the
outside. I would suggest that Orbaneja de Picos could originally have been
26
In medieval Castile, cf. Orbaneja del Castillo, a deserted Orbaneja, in the alfoz of Muñó
(G. Martínez Díez, Pueblos y alfoces burgaleses de la repoblación (Valladolid, 1987), p.309), as well
as the monastery of San Juan de Orbañanos (Orbanianos).
Page 172
Ordernr. 061297 P1
‘Orbaneja’ proper, as the main settlement of a homonymous area in which other

settlements (villae) existed. Unfortunately, we lack the necessary archaeological
evidence here to take the argument further.27
In tenth-century land transfers, properties are referred to the supralocal unit:
land ‘in the place called Orbaneja’ (in locum quod vocitant Orbanelia) (942); a
vineyard and a piece of land ‘in the Orbaneja valley’ (in valle de Orbanelia) (964
and 969); properties ‘in the surroundings of Orbaneja’ (in aditos de Orbanelia)
(953 and 984).28 The only extant eleventh-century transaction introduces the
village level: land ‘in the Orbaneja valley, in the place called Cardeñuela’ (in Valle
de Orbanelia, in locum quem vocitant Cardeniola).29 This is consistent with the
Ausín pattern, although here the supralocal level is better illustrated in earlier
texts, while the later evidence is scanty.
Grazing rights in the Orbaneja area yield a richer picture of supralocal
territoriality. In 963 Cardeña was granted the full villa of Orbaneja de Picos,30 in
the suburbium* of Burgos;31 the charter then goes on to award the monastery
grazing rights in a substantial area, limited by the River Arlanzón and the neigh-
bouring villages of Castañares, Villafría, and Rubena, up to the Atapuerca ridge
(Fig. 7.3). This means not only the whole Orbaneja area — as defined by topog-
raphy — but also part or all of the territory of other villages, such as Arlanzón or
Ibeas, which clearly did not belong to it. It seems that the grazing rights of the
inhabitants of Orbaneja reached beyond its boundaries — probably neighbouring
units had reciprocal rights in Orbaneja — and this can be seen as an attribute of
community membership, as eleventh-century texts make clear.
In 1039 King Fernando I donated to Cardeña the smaller monastery of San
Vicente ‘in the Orbaneja valley, in the urban territory of Burgos’ (in valle Orbanelie,
in suburbio Vurgos),32 which lay right on top of the Atapuerca ridge. Attached to
it were rights in pasture and water ‘down to the River Arlanzón’ (usque in flumine
Aslanzon), totally overlapping with the southern side of those assigned in 963 to
27
Compare Modúbar and Hontoria, above.
28
Cardeña, nos 39, 83, 116, 145, 192.
29
Cardeña, no. 267 (1050).
30
Donations of whole villages in Castile are a very different affair from mere land trans-
actions, as they involved seigneurial rights over the whole community as such. They are also much
less frequent, and tend to be issued by rulers or people from their immediate entourage.
31
Cardeña, no. 109 (963).
32
Cardeña, no. 233 (1039).
Page 173
Ordernr. 061297 P1
158 Julio Escalona
Orbaneja de Picos. The rest of the circuit is omitted, but it presumably coincided
with the boundaries of the Orbaneja valley itself, since San Vicente did not belong
to any particular village, but to the whole valley of Orbaneja.33 That very year,
King Fernando I also conferred upon Cardeña special exemptions in three villae
including Orbaneja de Picos,34 including those of a military nature — castle
repair, military expedition with the king or his delegates, providing men for recon-
naissance missions — and subjection to inspection by royal judicial officials, and
several others. As a result, the men of Orbaneja de Picos — Cardeña’s dependants
— were set apart from the ordinary network of public obligations and placed
exclusively under the abbot’s lordship. Yet, they retained their rights as members
of the wider community of Orbaneja. This situation triggered a dispute in 1073
over whether the men of Orbaneja de Picos were entitled to share the valley’s
pastures. In the lawsuit, thirteen local knights (infanzones*) spoke ‘for the whole
valley of Orbaneja’ — that is, four villas: Villa de Plano, Villa de Valle, Cardeñuela
and Villa de Domna Eilo35 — and were ultimately defeated by the monks.
Now, the 1073 lawsuit provides precisely the main factor lacking in the Ausín
material, namely, collective political action at a supralocal level. Between the tenth
and the eleventh centuries, ownership of agricultural land presumably became
increasingly confined to the component villages in the valley. However, uncul-
tivated land was managed jointly and shared by the people of the member settle-
ments. It was a part of a network of rights and obligations that can only be defined
as political. The men of Orbaneja de Picos had their grazing rights denied because
they had strayed beyond the common obligations. Their victory — actually
Cardeña’s — in court was a lethal blow to the very roots of the valley community,
and in fact supralocal articulation in Orbaneja fades out thereafter, at least in our
written evidence.
A brief review of the Ausín and Orbaneja areas yields a picture of territorial
change from larger, topographically consistent supralocal units to smaller village-
territories. It also reveals that this process was extremely uneven: in some areas it
took place earlier than in others. In some it went nearly totally unrecorded. Orba-
neja is remarkable for presenting us with supralocal community links in action at
a time when village territories were clearly the dominant pattern; in fact by then
33
This kind of ‘supralocal’ monastery is not infrequent, but nothing suggests that they could
exercise a pastoral role over the area. Cf. Bassett, this volume.
34
Cardeña, no. 234 (1039).
35
Cardeña, no. 340 (1073).
Page 174
Ordernr. 061297 P1
they were being used as the building blocks of the new parish system. We must
now turn our attention to the next territorial level: the administrative districts.
This will bring in the issues of spatial hierarchy and relationships with the overall
political system, which are crucial to a proper understanding of territoriality.
Hierarchy: Central Settlements and Administrative Districts
The other chief factor in Castile’s spatial complexity is the existence of central
settlements. I am not concerned here with the many different kinds of spatial
focus one can detect on the ground,36 but more specifically with higher-rank set-
tlements that played a role as a sort of capital of supralocal units and/or districts.
The study of early Castilian central settlements has been much neglected. Ad-
ministrative districts and their centres were long recognized, but it was their military
role that aroused the greatest attention.37 Most centres had some kind of fortress
to which the district’s free population owed military obligations. This led many
to think that defence was the only reason they were created, so the main compo-
nent of spatial hierarchization was seen as generated top-down by the Castilian
rulers in their ongoing struggle with the Muslims. Local society hardly played any
role other than accepting the imposition and fulfilling their obligations.
Since the mid-1980s a more complex vision has been gaining favour, starting
with a seminal paper by Carlos Estepa in which he defined the alfoces as resulting
from the interaction between existing territories and the emerging overall political
system.38 Since the late ninth century, not only governance, justice and defence,
but also the management of the counts’ patrimonial estates were based upon a
network of districts which was to endure until the late twelfth-century creation
of new, larger divisions: the so-called merindades menores*.39 Neither the alfoces
nor their centres were, however, created from scratch. Their raw material was a
36
Compare Davies’s conclusion to this volume.
37
Cf. Martínez Díez, Pueblos y alfoces, pp.10–12.
38
Estepa Díez, ‘El alfoz castellano’.
39
Álvarez Borge, Monarquía feudal. On the difficulties of separating which aspects derived
from the counts’ ruling functions, and which were due to their ‘private’ behaviour as estate
owners, landlords and patrons of clienteles, see J. Escalona, ‘Comunidades, territorios y poder
condal en la Castilla del Duero en el siglo X ’, Studia Historica: Historia Medieval, 18–19
(2000–01), 85–120.
Page 175
Ordernr. 061297 P1
160 Julio Escalona
web of small territories, deeply rooted in their local contexts, as further studies
suggest for Castile and other areas of the northern plateau.40
In what follows I explore two possible strategies by which this argument can
be pursued. The first consists of looking at district centres and considering their
relationship to previous territorial structures; the second consists of checking
whether district boundaries conform with the local and supralocal units discussed
above. In dealing with both aspects, a general distinction must be kept between
the larger districts, whose centres were the main nodes of the articulation of the
Castilian county, and the many smaller units that lay among them.
Central Settlements, Large and Small
The largest district in tenth-century Castile was Clunia. It is also a good example
of how long-term continuity combines with extreme changes. A pre-Roman town
(oppidum*) turned into Roman city, Clunia was the capital of a huge conventus
iuridicus* (administrative/judicial district) comprising all the eastern half of the
Duero basin and the corresponding northern lands down to the sea.41 The town
flourished until the third century, but its late Roman phase is characteristically
elusive.42 Being strongly dependent on the overall political system, it seems to have
declined in the fifth century to the point that the region’s episcopal see in the
Visigothic period was established in nearby Uxama (Burgo de Osma).43 Clunia
arguably retained some sort of regional focal role during the earlier Middle Ages,
although its character must have changed considerably, depending especially on
the scale of the overall political system. Tellingly, it reappears in early tenth-century
sources as a major stronghold in the upper Duero region and rises to become the
capital of the whole south of the Castilian county in the turbulent second half of
40
See, among others, for northern Castile: Martín Viso, Poblamiento y estructuras sociales, and
J. A. Lecanda Esteban, ‘Mijangos: arquitectura y ocupación visigoda en el norte de Burgos’, in II
Congreso de Arqueología Peninsular (Zamora, 1996), pp.415–34; and for the south: Escalona,
Sociedad y territorio, and F. Reyes, Población y sociedad en el Valle del Duero: Duratón y Riaza en
la alta Edad Media, siglos VI al XI (Madrid, 2002). Also, Martín Viso, Fragmentos del Leviatán.
41
M. C. García Merino, Población y poblamiento en la Hispania Romana: El Conventus
Cluniensis (Valladolid, 1975).
42
Cepas, this volume.
43
I. Martín Viso, ‘Organización episcopal y poder entre la Antigüedad Tardía y el Medioevo
(siglos V – XI): Las diócesis de Calahorra, Oca y Osma’, Iberia, 2 (1999), 151–90.
Page 176
Ordernr. 061297 P1
the century.44 In this later phase the boundaries of its huge district demonstrably
engulfed a number of smaller units, arguably formerly independent alfoces.45
Clunia may look unique because of its unparalleled importance in the Roman
period and in the late tenth to early eleventh century, but the centres of southern
Castile’s larger alfoces (Lara, Lerma, Muñó, Roa, Osma and others) can also often
be traced back to Roman towns, even to pre-Roman oppida. The mere fact that
many of them made it into the high medieval network speaks strongly for overall
continuity but it must be stressed that we can hardly reconstruct any part of their
evolution between the late Roman or Visigothic periods and the tenth century.
Moreover, if these centres date back to earlier times, we have almost no way of
knowing what their boundaries were before the evidence of the tenth-century
charters, and these often respond to demonstrably recent alterations, as is the case
of Clunia or Lara.
Burgos — Castile’s tenth-century political capital if it had one — can be
contrasted with Clunia. Despite its prominent position overlooking the River
Arlanzón and the presence there of a substantial prehistoric hillfort, finds of
Roman archaeological material are scanty and it has not been possible to relate it
to any node in the Roman itineraries. Instead it lay halfway between other sites,
such as Tritium (Monasterio de Rodilla) or Deobrigula (Tardajos). We cannot
ascertain why it succeeded in becoming Castile’s chief town in the absence of
more intense archaeological research,46 but its takeover in 882 by Count Diego
Rodríguez is a major landmark in the Castilian annals, and in the tenth century
it was clearly as close to a city as one could get in the county.47
Before alfoz became the standard term for district in the kingdom, the word
suburbium was commonly used both for large districts like Burgos as for smaller
44
Escalona, ‘Comunidades, territorios’. The medieval stronghold did not lie on the site of the
Roman city — itself built on an outstandingly impressive elevation — but on another hill nearby,
where the ruins of the medieval castle are still visible.
45
See further Escalona, ‘Comunidades, territorios’.
46
J. L. Uríbarri, J. M. Martínez González and I. Leis, Primeros asentamientos humanos en la
ciudad de Burgos (Burgos, 1987); T. Mañanes and J. M. Solana, Ciudades y vías romanas en la
Cuenca del Duero (Valladolid, 1985), pp.31–32; J. J. García González, ‘Del cerro al castillo: El
cerro de Burgos de la Antigüedad a la Edad Media’, Cuadernos Burgaleses de Historia Medieval,
2 (1995), 71–166.
47
Anales Castellanos I, ed. M. Gómez Moreno (Madrid, 1917), s.a. 882; C. Estepa Díez,
‘Burgos en el contexto del nacimiento de la ciudad medieval castellano-leonesa’, in La ciudad de
Burgos: Congreso de Historia de Burgos (Madrid, 1985), pp.23–55.
Page 177
Ordernr. 061297 P1
162 Julio Escalona
Photo 7.3. El Castillo, La Muela, and Peña de Lara, from Lara.
ones like Ausín, whose centre’s ‘urban’ character was more than dubious. This
suggests that former urban nodes and hinterlands may have played a role in
defining the new categories of territorial division, if not their actual nodes and
boundaries.
Secondary central settlements have attracted much less attention, and have an
even more incomplete archaeological record. Yet cases such as Ausín provide
interesting hints. Ausín’s fortress lay on a prominent hilltop overlooking the
valley (see Photo 7.2). Ramparts and enclosures are clearly recognizable, and the
site typically yields pottery from all of the late pre-Roman Iron Age, and from the
late Roman period.48 While, in the early Roman period, the clearest settlement
evidence comes from neighbouring lowland sites such as Hontoria de la Cantera
and Revilla del Campo, the old hillfort surely remained a strong symbolic presence
in the local landscape, although we cannot now be sure of the nature of its late
48
J. A. Abásolo Álvarez and I. Ruiz Vélez, Carta arqueológica de la provincia de Burgos: Partido
judicial de Burgos (Burgos, 1977), pp.19–21.
Page 178
Ordernr. 061297 P1
Roman phase.49 In the absence of excavations, everything suggests that the early
medieval settlement did not occupy the hilltop, but was dispersed in a linear
pattern along the river, eventually to become nucleated around three foci, each
with its own church (another pointer to a higher-rank settlement). By the tenth
century — although we do not know from how long before — a castle had been
built on the hilltop, which was probably redundant by the early thirteenth century
when a church was built, which preserved the site’s symbolic relevance down to
the present day.50
Ausín’s case is typical of small alfoz centres. Most had a recognizable fortress,
very often placed upon an earlier hilltop site*. Surely there is an argument for
long-term continuity there but this should be qualified by other facts: a) while
ancient sites demonstrably lie under or close to most tenth-century central settle-
ments, others of comparable or greater size did not make it into the Middle Ages;
b) the most common pattern is that of a site with late prehistoric material, then
a gap for the early Roman period (sometimes starting already in the so-called
Celtiberic-phase), then a late Roman reoccupation whose end is typically hard to
determine, then an early medieval fortress. Overall, it seems that those sites
repeatedly rose as dominant in the neighbouring landscape, but the process is far
from simple.51
Up to this point, it seems clear that the central fortresses of alfoces were com-
monly placed on sites that often had a local relevance and a long-running history.
Yet, were those castles erected or renewed because they were chosen by the Cas-
tilian rulers as alfoz centres or were they chosen to be so because they were already
higher-rank sites in the local landscape, whether they had a fortress or not? We
can hardly answer that question at all, at least not by considering the sites alone.
At this point, looking at district boundaries will certainly not solve the problem,
but it can provide valuable data.
While the boundaries of large districts like Clunia extend beyond and across
topographical limits, smaller territories are potentially more likely to keep to the
constraints of watercourses, watersheds and other physical landmarks (which of
49
Martín Viso, this volume, on hilltop site reoccupation; and Halsall and Cepas, this volume,
on community redefinition following the ending of the Roman patterns of socio-territorial
domination.
50
J. Escalona ‘Poblamiento y organización territorial en el sector oriental de la cuenca del
Duero en la Alta Edad Media’, in III Congreso de Arqueología Medieval Española, 2 vols (Oviedo,
1992), II, 448–55.
51
Discussion in Escalona, Sociedad y territorio, pp.48–55, and Martín Viso, this volume.
Page 179
Ordernr. 061297 P1
164 Julio Escalona
course entails a clear risk of determinism). This does not mean, though, that they
always did; indeed many alterations are demonstrated.
Take the small valley of the River Mataviejas (whose medieval name was Ura).
This is a narrow east–west corridor interrupted in its middle sector by steep,
curving slopes, a major obstacle to communication along the valley. In the tenth
and eleventh centuries the valley was divided into two alfoces: up-river, Tabladillo
— whose centre lay on what was arguably a first-century nucleated settlement
(vicus*) — and, down-river, Ura, a district homonymous with the river and with
its centre, below a hilltop site where late Roman material has been collected. The
two alfoces are clearly defined by the topography on both sides, but, while their
northern boundaries closely match watersheds, their southern limits extend
beyond the watershed and eat into other neighbouring hydrographic units. Most
of the component villages of the alfoz of Tabladillo, as deduced from a list in an
1125 charter,52 kept to the valley limits, yet a land tract protruded southwards to
include the village of Doña Santos, clearly alien to the valley. This arrangement
was much earlier than 1125: Talamanquilla — further south from Doña Santos
— belonged to Tabladillo until it was granted to Covarrubias Abbey in 978.53
Therefore, Tabladillo’s southern expansion dated back at least to the military
difficulties of the tenth century that triggered the reorganization of the frontier
sector dominated by Clunia.54 This may also shed some light on the origins of
similar features in neighbouring Ura, whose district also shows a south/south-
west expansion beyond watersheds in the Nebreda-Cebrecos sector, which we
cannot explore further through written sources.
Phenomena of this kind are better recorded in the Ausín district (Fig. 7.2). In
the tenth century it comprised several units which eventually split into two or
more villages each (Hontoria, Revilla, the Modúbares, the Cubillos). Yet, before
that situation came about, their internal consistency was also affected by changes
in district boundaries. Take the Modúbar valley. While place-names show it as a
unit, its western half was captured by the all-important alfoz of Burgos, as was the
case with other neighbouring territories.55 Much later, the now deserted village
52
Documentación del monasterio de Santo Domingo de Silos (954–1254), ed. M. C. Vivancos
(Burgos, 1988), nos 37–38.
53
Cartulario del Infantado de Covarrubias, ed. L. Serrano (Valladolid, 1907), nos 7–8.
54
Escalona, ‘Comunidades, territorios’.
55
J. Escalona, ‘Acerca de la territorialidad en la Castilla altomedieval: Tres casos significativos’,
in Historia social, pensamiento historiográfico y Edad Media, ed. I. Loring García (Madrid, 1997),
pp.217–44.
Page 180
Ordernr. 061297 P1
of Tabladillo — not the same as the aforementioned alfoz of Tabladillo —

belonged to Ausín in 1011, but was captured thereafter by neighbouring Lara, its
name changing to Tabladillo de Lara.56 A third occurrence — Lara’s absorption
of the uppermost settlement in the Cubillo valley, whose name changed to
Cubillejo de Lara — has left little more than traces in the place-names.
The fact that, over a long time-span, alfoz boundaries cut across smaller terri-
torial units regardless of topographic and/or toponymic unity not only shows that
there was a degree of autonomy between local and higher political levels. It also
suggests that there was a pre-existing territorial structure which was partially
preserved and reused, and partially modified in the course of time.
Scale Change and Territories: Where Is the Community?
The latter part of this chapter has dealt with administrative districts, which, like
ecclesiastical divisions,57 belonged to large-scale networks of power. This raises the
major issue: the relationship between the top-down territorial units portrayed in
our sources and the webs of community relationships that we want to investigate.
We are as yet far from having fully satisfactory answers to this question. Suffice it
to say that it is complex enough to rule out any simple explanation.
In the cases I have discussed, and despite the increasing stress upon the village-
community, community bonds seem to be better defined as a complex set of over-
lapping, multifaceted relationships, running from the very local to the supralocal.
Residence is the first obvious sphere, but we still need much more archaeological
research if we are to understand when and how the dominant pattern of discrete
nucleated settlements came into being. Also, the evidence discussed suggests that
property patterns and economic activities may well have worked at a supralocal
scale (Orbaneja, Modúbar, the Cubillos) before they became more and more
confined to the individual villages within them. Patterns of access to uncultivated
land and other resources such as water, woodland or mining seem to have worked
at a similar middle-scale. The larger, top-down designed districts were not suitable
for that. For all we know, neither Burgos nor Clunia were grazing-rights units,
and Lara only laid that claim at a very late date. Orbaneja was clearly this sort of
territory as late as 1073, and there are hints that in other small territories such as
56
Cartulario de San Pedro de Arlanza, ed. Serrano, no. 111.
57
Bassett, this volume.
Page 181
Ordernr. 061297 P1
166 Julio Escalona
Tabladillo or Valdelaguna, grazing rights worked at an alfoz scale and were

relevant to community membership.58
I hope to have shown that southern Castilian tenth- and eleventh-century
territorial structures were more complex and dynamic than the established views
would have us think. Individual settlements belonged to several layers of overlap-
ping territoriality: villages, larger valley-like settlement units, a centre’s hinterland,
the county, and the kingdom. Moreover, the whole system was changing at a rapid
pace, determined by the inclusion of a large number of pre-existing building-
blocks into Castile’s increasingly integrated political framework. Whether ab-
sorbed into larger districts or turned into districts themselves, the early units gave
way to a system whose greater emphasis was on the local articulation of central
power,59 that is, of the main mechanisms by which the local population interacted
with an ever more distant political system: justice, tribute-collecting, labour-
services, military obligations, and so on. In the long run, this favoured a threefold
pattern.60 Households were arranged in local communities of village/parish size,
in turn grouped into larger units that were more the administrative tools of top-
down government and less a territorial expression of the complexity of local
society. Households and villages were suitable as units for the extraction of both
taxes and lordly revenues, which ultimately depended upon domination over
peasant households and ownership of agricultural land. The twelfth-century
formation of the new division in merindades menores, only relatively dependent
on the earlier alfoces, was a major landmark in this separation between the village
and the district. The earlier supralocal territories seem to have largely faded out,
but left traces in the form of grazing rights shared between groups of villages, or
supralocal patterns of military obligations or tribute payments, into the late
medieval and early modern periods.
58
J. Escalona, ‘Jerarquización social y organización del espacio: Bosques y pastizales en la Sierra
de Burgos’, in Los rebaños de Gerión: Pastores y trashumancia en Iberia antigua y medieval, ed.
J. Gómez-Pantoja, Collection de la Casa de Velázquez, 73 (Madrid, 2001), pp.109–37.
59
S. Castellanos and I. Martín Viso, ‘The Local Articulation of Central Power in the North
of the Iberian Peninsula (500–1000)’, Early Medieval Europe, 13 (2005), 1–42.
60
J. A. García de Cortázar and E. Peña Bocos, ‘De alfoces, aldeas y solares en la Castilla de los
siglos IX al XI: ¿una formalización “feudal” del espacio?’, in Miscellania en Homenatge al P. Agustí
Altisent (Tarragona, 1991), pp.183–202.
Page 182
Ordernr. 061297 P1
C ENTRAL P LACES AND THE T ERRITORIAL

O RGANIZATION OF C OMMUNITIES:
T HE O CCUPATION OF H ILLTOP S ITES
IN E ARLY M EDIEVAL N ORTHERN C ASTILE
Iñaki Martín Viso
The Occupation of Hilltop Sites: Ancient Roots, New Phases
I
n spite of the introduction of new methods and theories in Castilian histori-
ography on the early Middle Ages,1 the lack of contemporary sources makes
analysis difficult and different explanations exist, many of them highly specu-
lative. One possible way of transcending this problem is through the study of the
social evolution of settlements as part of the transformations of late antiquity and
the early Middle Ages. In this context, I want to focus my analysis on the
occupation of hilltop sites*, whose occupation is revealed in our data from the
fifth century onwards.2 Ninth- and tenth-century written documents and chroni-
cles frequently refer to forts, named castra or castella*, in different districts and
archaeology has made clear the vitality of these hilltop sites in some areas of the
Iberian peninsula.3 In describing the occupation of these sites, two essential
1
See Escalona, this volume.
2
A clear example is the mention of the castella tutiora, in which the population took shelter
from the Suevic incursions, Hydatius, Chronicon, ed. J. Campos (Salamanca, 1984), VI.439–45.
3
See A. Barrios García and I. Martín Viso, ‘Reflexiones sobre el poblamiento rural altome-
dieval en el Norte de la Península Ibérica’, Studia Historica: Historia Medieval, 18–19 (2000–01),
53–83; J. López Quiroga, ‘Fluctuaciones del poblamiento y hábitat “fortificado” de Altura en el
Noroeste de la Península Ibérica (ss. V – IX )’, in Mil Anos de Fortificaçôes na Península Ibérica e no
Magreb (500–1500) (Lisbon, 2001), pp.83–91.
Page 183
Ordernr. 061297 P1
168 Iñaki Martín Viso
factors must be underlined. On the one hand, not all occupation was similar,
meaning that the places under study had different histories: different functions,
varied degrees of intensity of occupation, ruptures and continuities should be
identified on each site. On the other hand, the Iberian landscape, specifically its
northern and central regions, is very heterogeneous and so in some areas there was
no general process of occupation of hilltop sites.4
Hilltop sites became very important in the territory that became the County of
Castile from the first third of the tenth century. The name Castile, which began to
be common from the second half of the ninth century, comes from Latin castella and
refers to the abundance of fortifications. Many recent analyses have pointed out
the role of these places as foci for the socio-political articulation of early medieval
Castile.5 I want to discuss some problems and make some remarks on that issue
through the study of one specific area: northern Castile, the region around the
high basins of the rivers Ebro and Pisuerga, now part of the provinces of Palencia,
Burgos, Cantabria and La Rioja (Fig. 8.1). This zone of approximately 5000 km2
was identified, from the tenth century, with the district of Castella Vetula (Old
Castile). It is very fragmented geographically as well as being very heterogeneous
in terms of its early medieval landscape and social evolution.
The most usual explanation for the occupation of hilltop sites postulates their
abandonment after the Roman conquest and their later reoccupation from the
fifth century onwards, on the basis of the clear hiatus in the archaeological
record.6 Their reactivation would result from the increase in political instability,
although their occupation was less intense than in the pre-Roman period.7 In the
4
For example, the Cantabrian areas. See I. García Camino, Arqueología y Poblamiento en
Bizkaia (siglos VI– XII): La Configuración de la Sociedad Feudal (Bilbao, 2002) and J. A. Gutiérrez
González, ed., Peñaferruz (Gijón): El Castillo de Curiel y su Territorio (Gijón, 2003).
5
See J. Escalona Monge, Sociedad y territorio en la Alta Edad Media Castellana: La formación
del alfoz de Lara, British Archaeological Reports, International Series, 1079 (Oxford, 2002);
R . Vázquez, ‘Castros, castillos y torres en la organización social del espacio en Castilla: el espacio
del Arlanza al Duero (siglos IX a XIII)’, in Del Cantábrico al Duero: Trece Estudios sobre
Organización Social del Espacio en los siglos VIII a XIII, ed. J. Á. García de Cortázar (Santander,
1999), pp.351–73, and I. Martín Viso, Poblamiento y estructuras sociales en el norte de la Península
Ibérica (siglos VI– XIII) (Salamanca, 2000).
6
See J. A. Gutiérrez González, Fortificaciones y Feudalismo en el Origen y Formación del Reino
Leonés (siglos IX – XIII) (Valladolid, 1995).
7
A. Domínguez Bolaños and J. Nuño González, ‘Reflexiones sobre sistemas defensivos
tardoantiguos en la Cuenca del Duero: a propósito de la muralla de “El Cristo de San Esteban”,
Muelas del Pan (Zamora)’, in La Hispania de Teodosio, ed. R . Teja and C. Pérez, 2 vols
(Salamanca, 1997), II, 435–50.
Page 184
Ordernr. 061297 P1
CENTRAL PLACES AND THE TERRITORIAL ORGANIZATION 169
Figure 8.1. Hilltop sites of Northern Castile (sixth–eleventh centuries).
Page 185
Ordernr. 061297 P1
Castilian case, the rupture between pre-Roman and medieval strata appears on
many sites. Thus, Monte Cildá (Olleros del Pisuerga), doubtless to be identified
as the main focus of the late antique civitas* of Mave, was reused from the fifth to
eighth centuries. Some refortification was carried out after settlement was moved
in the Roman period to the plain around the present-day village of Santa María
de Mave.8 However, tenth- and eleventh-century written documents reveal
numerous examples of hilltop sites as functioning elements within the settlement
pattern. The documents describing the endowment of the monastery of San
Salvador de Oña at its foundation in 1011 mention several, such as Castello Dei
(probably Medina de Pomar), Santa Gadea del Cid or Alfania.9 Moreover, many
of the place-names of these sites seem to be pre-Roman (note the use of the suffix
-on in the original forms of names such as Siero, Berbeia or Alfania) and others
describe a defensive structure (muratiello, castriello). Although this is not a deci-
sive argument, it is very strange that the pre-Roman names of these sites, which
were completely abandoned without developing any local function, should have
been preserved for more than four hundred years. Furthermore, the new and
permanent occupation of some hilltop sites throughout our period must have
involved profound transformations in local economic organization, and these
cannot have been brought about very quickly. Consequently, the theory of reoc-
cupation cannot provide a general explanation for the occupation of hilltops,
although it would have applied in some cases. A possible explanation is that some
of these sites could have survived with certain important transformations as part
of the Roman network.
The archaeology carried out to date provides poor information but there are
some sites that yield interesting, though problematic, data. One is Siero, a hill fort
on the confluence of the rivers Rudrón and Ebro, formed by a residential area and
probably an old fortification located on the summit, where a hermitage devoted
to Saints Centola and Elena now stands. Surveys carried out in that area have
discovered an Iron Age phase followed by a Roman one but the most usual
evidence found is a kind of smooth or grooved wheel-turned painted pottery,
dated to the early Middle Ages but without a clear chronology.10 The presence of
8
L. Hernández Guerra and L. Sagredo San Antonio, La Romanización del territorio de la
actual provincia de Palencia (Valladolid, 1998), pp.44–54 and 108–09.
9
Colección Diplomática de los Condes de Castilla, ed. M. Zabalza Duque (Salamanca, 1998),
doc. 64.
10
R . Bohigas Roldán and others, ‘Carta Arqueológica de la Provincia de Burgos: Partidos
Judiciales de Sedano y Villarcayo’, Kobie, 14 (1984), 7–91 (pp.21–22 and 28); J. Campillo Cueva,
Page 186
Ordernr. 061297 P1
an eighth- or ninth-century inscription, still extant in a wall of the hermitage,

seems to underline the use of this place during that period, although its present
location is not the original one.11 According to written records, this place gave its
name to a territorial district or alfoz* during the tenth and eleventh centuries, and
it might have been the centre of a supralocal unit*.12 Another similar case is
Berbeia, located on one of the promontories near Bachicabo Peak, just above
Barrio de Valdegovía, with visual control of the valley of the Omecillo. Excava-
tions carried out there revealed a cultural sequence that started in the second Iron
Age and continued, although with clear contraction, through the Roman period.
In the Middle Ages, however, this place was used again, with an occupation de-
fined as ‘light’ compared with the pre-Roman phase, and detected on the basis of
some wheel-turned pottery. The settlement was divided into two sections, a resi-
dential area, with the remains of buildings with stone foundations, and a defensive
one, situated higher up.13 Written documents reveal its presence as a populated
site in the early medieval period, and it is still mentioned in 1123.14
A pronounced hiatus during the early Roman period was apparently common.
A very rigid view of the archaeological data compels us to consider a sudden break-
down in the use of hill forts during the Roman era followed by later reoccupation.
However Roman phases were identified only by the presence of sherds of pottery
known as terra sigillata. Some occupation from the Roman period could, there-
fore, be hidden, particularly in peripheral zones, and only revealed in the late
Roman centuries thanks to late Hispanic sigillata (Terra Sigillata Hispanica
tardía). I want to propose another view, based on the different degrees of intensity
in the use of hilltop sites during successive periods. A territorial hierarchy was
formed, focussed on hill forts with separate residential and defensive functions in
two sections and with a subordinate territory round about. Roman domination
undoubtedly favoured the development of new types of settlement linked to
‘Los despoblados medievales de la Honor de Sedano (Burgos)’, Kobie, 24 (1997), 125–68

(pp.152–53).
11
S. Andrés Ordax, ‘Arte Burgalés de la alta edad media’, in II Jornadas Burgalesas de Historia:
Burgos en la Alta Edad Media (Burgos, 1991), pp.125–40 (pp.135–36).
12
Becerro Gótico de Cardeña, ed. L. Serrano (Valladolid, 1910), doc. 270; I. Martín Viso,
‘Monasterios y Poder Aristocráticos en Castilla en el siglo XI’, Brocar, 20 (1996), 91–133 (doc. 6).
13
J. A. Algorreta and others, ‘Castro de Berbeia (Barrio-Álava): Memoria de excavaciones,
Campaña de 1972’, Estudios de Arqueología Alavesa, 8 (1975), 221–92.
14
Colección Diplomática Condes de Castilla, ed. Zabalza Duque, docs 30 and 73; Los
Cartularios Gótico y Galicano de Santa María de Valpuesta (1090–1140), ed. S. Ruiz de Loizaga
(Vitoria, 1995), Gótico, doc. 136.
Page 187
Ordernr. 061297 P1
structures of ownership and power that did not exist before, to the profit of some
local aristocratic groups, as in the villa-estates* of Las Ermitas (Espejo) or those
of the Losa Valley.15 The intensity of the use of hilltop sites probably decreased,
but they continued to form part of the Roman rural landscape. The disintegration
of the Roman system promoted the renewed influence of local aristocratic groups,
diluting the hierarchy established by the Romans. Later, the Visigothic realm
needed to negotiate with local authorities, in order to strengthen its political
dominion. In this context, each community could obtain greater autonomy and
so used the hilltop sites more intensively. Thus, a sequence generally dated as early
medieval can be detected. The function of these places was above all political and
they worked as a territorial representation of communities’ autonomy, and espe-
cially as refuges. The key lies in considering the hilltop site not as an isolated ele-
ment, but as part of a system: its creation cannot be understood outside a pattern
of exploitation of resources founded in strong community autonomy. It was a
dynamic system and changed down the centuries, with successive transformations
that did not break the general pattern and always kept the hill fort, although the
latter lost some of its previous functions during the Roman Period.16
Peasant and Central Power Initiatives
The location of many hilltop sites in northern Castile reflects an interest in mak-
ing complementary use of lowland and upland areas. The choice of this pattern
was connected with small-scale transhumance or with the desire to control the
richest agrarian land, and so the site would be located outside these areas. The
hilltop sites generally preferred marginal locations close to local routes, but not to
Roman roads, and do not reveal any large-scale architectural works. These features
imply that they were spaces created by very cohesive communities with strong
internal autonomy.
A good example of this type of territorial organization is Valderredible, a valley
situated in present-day Cantabria, close to the source of the Ebro. One of the
Valderredible’s hilltop sites is Ruanales, a place quoted in a document of 962 that
15
I. Filloy Nieva, ‘Yacimiento de Las Ermitas (Espejo)’, Arkeoikuska (1995), 295–302; J. A.
Abásolo Álvarez and F. Pérez, ‘Excavaciones en Salinas de Rosío (Burgos)’, Noticiario Arqueológico
Hispano, 24 (1985), 159–264.
16
I. Martín Viso, ‘Pervivencia y transformación de los sistemas castrales en la formación del
feudalismo en la Castilla del Ebro’, in Comunidades locales y poderes feudales en la Edad Media,
coord. I. Álvarez Borge (Logroño, 2001), pp.255–88.
Page 188
Ordernr. 061297 P1
refers to the presence of lands and mills ‘in Rabanales’ and perhaps refers to a
territory, although this is not certain.17 Its location seems to be the current village
of the same name, situated on a hill 970 m high, controlling an area of pasture
near Mount Ijedo as well as the basin of the river Panero.18 There are no remains
of fortifications at this site but, since they could have been of timber, they would
probably have been destroyed by the present residential area. Little is known
about the early medieval settlement pattern; the only contemporary remains in
the area — although problematic — are cemeteries of graves cut into the rock in
the neighbourhood. They probably were linked to a dispersed pattern, but we
need more information and research on these sites. The definition of this territory
is best explained by the exploitation of summer pastures on Mount Ijedo, together
with the lowland ones of the Panero basin (Fig. 8.2).19
A similar site is Castrillo de Valdelomar, the centre of the district of Valdelo-
mar, a stretch of valley in the West Valderredible. This site is located at the top of
a hill, from where there is visual control of the plain of the Mardancho stream,
close to Mount Ijedo and the bleak plateau Lora. Although there is little docu-
mentary evidence for Castrillo de Valdelomar, previous occupation can be
accepted on the basis of the existence of a cemetery of graves cut into the rock on
the hill where the present-day parish* church stands.20 A local route, traversing
the valley horizontally, had great influence in connecting the places of the district.
Therefore, the main impulse was the control of the natural resources of an area
which preserved strong communal ownership and whose origin was probably
linked to this territorial organization.21 The third hilltop site is Alfania, in the
middle of Valderredible, whose location can be identified with the so-called
Ermita del Monte, on the foothills of Mount Ijedo, 2 km away from the present-
day village of Villanueva la Nía. This site and its district, defined as an alfoz, were
mentioned in the first half of the eleventh century.22 Its location seems to be
17
Colección Diplomática del Monasterio de Sahagún (siglos IX y X ), ed. J. M. Mínguez (León,
1976), doc. 207.
18
R . Bohigas Roldán, Yacimientos Arqueológicos Medievales del Sector Central de la Montaña
Cantábrica (Santander, 1986), pp.190–91 and 196.
19
This last area was part of the alfoz of Bricia after the division of Castile between García III
and Fernando I in the middle of the eleventh century. Burgos, doc. 178.
20
Bohigas Roldán, Yacimientos, pp.188–89.
21
See J. Ortega Valcárcel, La Cantabria Rural: Sobre la Montaña (Santander, 1987),
pp.15–16.
22
Colección Diplomática Condes de Castilla, ed. Zabalza Duque, docs 64 and 75; Colección
Diplomática San Salvador de Oña (822–1284), ed. J. del Álamo (Madrid, 1950), doc. 35.
Page 189
Ordernr. 061297 P1
Figure 8.2. Territorial organization and hilltop sites

in the valley of Valderredible.
related to the domination of some areas close to the Ebro, perhaps with a dis-
persed settlement pattern. Some of the most important information about this
territory comes from a document of 1014, in which the count of Castile promised
to defend the use of communal lands by the inhabitants of Berzosa and Alfania,
represented by some local leaders (‘omes de concilio’).23 That piece of land was the
economic basis of the system ruled from the hilltop site (Fig. 8.3).
The study of Valderredible shows how these hilltop sites shaped territorial
structures larger than the villages. This kind of articulation was probably devel-
oped as a dominant factor in the early medieval Castilian landscape.24 They
existed in the tenth and eleventh centuries but perhaps also in the previous period
too, like the rock-cut graves and the pre-Roman place-names of the central sites
23
Colección Diplomática Condes de Castilla, ed. Zabalza Duque, doc. 75.
24
See Escalona, this volume. I. Álvarez Borge, Monarquía feudal y organización territorial:
Alfoces y merindades en Castilla (siglos X – XIV ) (Madrid, 1993).
Page 190
Ordernr. 061297 P1
Figure 8.3. The territory of Alfania-Paredes Rubias.
mentioned above. Nevertheless, not all the valley was organized into territories
around these hilltop sites and some areas were excluded, because another pattern
of territorial articulation emerged during these centuries. The most evident sign
is the villa-estate of Santa María de Hito, whose residential area was used as a
cemetery after the Roman era.25 The establishment of the nearby monastery of
San Martín de Elines, which retains some remains of an early medieval ‘mozara-
bic*’ church,26 would have been closely connected with the transformation of the
social basis of the aristocracy in the post-Roman period. The local elite invested
in this religious centre to strengthen its legitimacy and safeguard extensive,
indivisible ownership.27 This zone must be identified with the territory of Val de
Ripa Yble, a space without a central hilltop site.28
25
R . Gimeno García-Lomas, ‘El conjunto de cerámicas medievales de Santa María de Hito’,
in I Congreso de Arqueología Medieval Española, 5 vols (Zaragoza, 1985), II, 385–401.
26
Bohigas Roldán, Yacimientos, pp.199–200.
27
See Halsall, this volume.
28
See Burgos, doc. 20.
Page 191
Ordernr. 061297 P1
In spite of the priority of peasant or community initiative, the central

authorities sometimes promoted the occupation of hilltop sites. One of their main
features was their location in certain geo-strategically important areas, like gorges
or places near important routes. However, the aim was not exclusively to guar-
antee the use of military routes but also to make their domination of the region
effective through ‘islands of authority’, especially over the most powerful agrarian
zones. The genesis of some of those places can be traced back to the Roman
period, although they undoubtedly saw more intensive use during the period
between the sixth and eleventh centuries. The most eminent of this kind of site
in northern Castile is Tedeja. It is located on the foothills of Sierra de la Tesla,
from which one could control the wide plain between Villarcayo and Medina de
Pomar — one of the main agrarian areas of the region — and the La Horadada
pass, a narrow gorge of the Ebro. Tedeja was initially a Roman turris (tower) but
it was turned into a fortified nucleus in the Visigothic period (sixth–seventh
centuries). The considerable dimensions of this site and the complex works
carried out, with the raising of towers and very solid stone walls, have to be
emphasized because they are specific characteristics of Tedeja.29 It was probably
used as a focus from where Visigothic authority could have been extended in the
area, and the most eloquent traces of Visigothic presence in northern Castile have
been found in precisely this place. The consecration of the church of Santa María
de Mijangos, at the bottom of Tedeja, by Bishop Asterio of Oca at the end of the
sixth century implies that the place also served as the region’s ecclesiastical focus.
The residential area was not nucleated on the hilltop site, but developed on the
plain, close to the river, in Vallejo de Santillán, where a cemetery and some
buildings have been found (Fig. 8.4).30
The pattern of Tedeja shows some features of the hilltop sites created by the
central power: the priority placed on a geo-strategic position, the clear difference
between a large political and military area and a residential one, the consolidation
of spacious walled enclosures and the use of complex techniques of fortification.
The shaping of these ‘islands of authority’ was subordinated to the raising of
29
R . Bohigas Roldán and others, ‘Evolución de las formas y funciones de la arquitectura en
el Norte de Hispania: el caso de Tedeja’, in 3º Congresso de Arqueología Peninsular, 8 vols
(Oporto, 2001), VI, 555–68; J. Á. Lecanda Esteban, ‘Mijangos: la aportación de la epigrafía y el
análisis arqueológico al conocimiento de la transición a la alta edad media en Castilla’, in Visigodos
y Omeyas: Un Debate entre la Antigüedad Tardía y la Alta Edad Media, ed. L. Caballero and
P. Mateos (Madrid, 2000), pp.181–206 (pp.194–97).
30
Lecanda Esteban, ‘Mijangos’, pp.197–99.
Page 192
Figure 8.4. The area of Tedeja.
tribute and the establishment of political power structures into which the local
leaders were undoubtedly inserted. A similar case was Buradón, a site of Roman
origin formed by a fort, probably used between the fifth and tenth centuries, and
by a settlement on the slope. This place controlled the gorge of ‘Conchas de Haro’
and it was close to the basin of Miranda, a very rich agrarian plain. Archaeological
work has revealed a military structure and a church, perhaps raised in the fifth
century, which underwent enlargement during the early Middle Ages.31 Both
elements were connected with the area’s political and religious frameworks. On
31
M. Unzueta and A. Martínez, ‘Proyecto de variante y túnel entre las Conchas de Haro y
el Cruce de Briñas’, in Arqueología de Urgencia en Álava, 1989–1993 (Vitoria, 1994), pp.43–60.
Page 193
Ordernr. 061297 P1
the other hand, the fact that the houses were built in stone would mark it out as
a prestigious site. This site must be linked to Bilibio, a place mentioned in the
seventh century as a fort (castellum Bilibium),32 whose supposed remains have
been located on the other side of the pass, where the hermitage of San Felices de
Bilibio now stands. A stretch of wall made of regular stone blocks, of a type
unknown in the communal structures, is preserved in this small site. One possible
hypothesis would be that this last site was a fortified strong-point linked to
Buradón, which was the axis of the territory called Bilibio.
Both examples allow us to assert that the Visigothic kingdom of Toledo was
able to form ‘islands of authority’ from which its power was extended in northern
Castile, though becoming weaker the further it radiated out from the centre. But
they are the only cases — alongside Mave — of this kind of hilltop site in northern
Castile. That pattern was later the key to the articulation of socio-political higher-
rank polities during the eighth and ninth centuries, when any centralized power
disappeared in the meseta. Tedeja seems to have been the main nucleus of one of
the polities identified in 86533 that would later give rise to the political formation
based from 930 around the figure of Fernán González. On the other hand, the
same Fernán González was mentioned in 964 as count in Buradón,34 which would
reflect that place’s importance as an axis of political dominion in northern Castile.
However, the absence of central authority combined with the permanence of the
socio-political basis of local leadership favoured the creation of new places of
power focussed on certain hilltop sites. The most significant was Lantarón, seat
of one of the main Castilian counties before unification.35 Its location is still
unknown, but it was probably on the site where the hermitage of San Martín de
Lantarón now stands. If this identification is correct, its location repeated the use
of a spot overlooking a gorge of the Ebro at a strategic point, thanks to the
confluence of the routes that cross the valleys of Tobalina and Valdegovía.
32
Sancti Braulonis Caesaragustanis Episcopi, Vita Sancti Emiliani, ed. L. Vázquez de Parga
(Madrid, 1943), section 9, lines 3–4.
33
Ibn Idari, Historia de Al-Andalus, ed. F. Fernández Gónzález (Málaga, 1999), p.135. The
so-called lord (sahib) of Misanika (Mijangos, on the foot of Tedeja) was quoted in this text,
referring to an Arab campaign against the Castilian region.
34
Cartulario de San Millán de la Cogolla (759–1076), ed. A. Ubieto Arteta (Valencia, 1976),
doc. 85.
35
See I. Martín Viso, ‘Poder político y estructura social en la Castilla altomedieval: el condado
de Lantarón (siglos VIII– XI)’, in Los Espacios de Poder en la España Medieval, coord. J. I. de la
Iglesia Duarte (Logroño, 2002), pp.533–52.
Page 194
Ordernr. 061297 P1
Nevertheless, there are no remains of fortifications; if they existed, they have been
demolished. The aim was the control of the rich agrarian areas of Valdegovía and
Miranda, in which some minor territories focussed on hilltop sites endured.36
The Transformations From the Eighth to the Eleventh Centuries
Social, political and economic organization between the sixth and eleventh cen-
turies in the Duero basin was based on very different cells of local power, such as
hill forts, cities, monasteries or villa-estates. Each one worked as an arena wherein
any real influence of central power had to be negotiated. But the system experi-
enced a great change during the eighth century, due to the collapse of central
authority. The new status quo is explained by the disintegration of the Visigothic
kingdom after the Arab invasion of 711, the low interest of the Umayyads in the
direct control of the northern Iberian plateau and the inability of the regional
aristocracies to restore a central political apparatus. The local units, well-adapted
to the particular needs of their communities, consolidated their role in a frag-
mented world.37 Many small-scale leaderships emerged throughout this region,
but none achieved hegemony. The formation of some polities was a long process,
culminating in this part of the meseta in the crystallization of the County of
Castile around 930. A determinant factor was the promotion of some aristocratic
groups, distinguished by a strong military ethos and able to be inserted into a new
large-scale system, working as the dominant social group.38 The supralocal units
were transformed into the basic axis of the configuration of aristocratic power,
because they were, at the lower, smaller-scale, level, the arena for economic, social,
political and religious organization. Many were focussed on hilltop sites, although
there are outstanding exceptions. These transformations should be connected
with the progressive feudalization of Castilian society, which produced very dif-
ferent effects. I want to draw attention to one of them, the ‘aristocratization’ of
the hilltop settlements and the gradual loss of their function as a main focus of the
36
See Martín Viso, ‘Poder político’, pp.538–39.
37
S. Castellanos and I. Martín Viso, ‘The Local Articulation of Central Power in the North
of the Iberian Peninsula (500–1000)’, Early Medieval Europe, 13 (2005), 1–42; Escalona Monge,
Sociedad y territorio, pp.58–72; I. Martín Viso, Fragmentos del Leviatán: La articulación política
del espacio zamorano en la alta edad media (Zamora, 2002), pp.39–65.
38
I. Álvarez Borge, ‘Estructuras de poder en castilla en la alta edad media: señores, siervos,
vasallos’, in Señores, Siervos, Vasallos en la Edad Media (Pamplona, 2002), pp.269–308.
Page 195
Ordernr. 061297 P1
spatial hierarchy. They were an instrument through which the local leaders’ mili-
tary monopoly could be consolidated, and the political and economic resources
of communities controlled — due to the centralization of military duties in the
fortress — and, in general, through which other social groups could be dominated.
The passage from ‘peasant castles’ to ‘aristocratic fortress’ was very simple in
the case of the great nuclei created by the central authorities. Tedeja is a good
example, because it was one of the early counties of the ninth century and later a
seat of Castilian power.39 Excavations show the persistence of a strong defensive
structure and the remodelling of the adjoining ancient church of Santa María
(probably abandoned in the tenth century), resembling the settlement of Vallejo
de Santillán.40 This area was very important in the articulation of large-scale
power before the crystallization of the County of Castile. It probably lay at the
origin of the district of Castella Vetula, in which other lesser-rank sub-systems
were inserted.41 Lantarón and Buradón had a similar evolution and they were the
chief places of important districts from which the Castilian county was formed
in this area. However, they experienced a clear decline from the end of the tenth
century. They were replaced by new sites (like Término-Santa Gadea del Cid, to
the detriment of Lantarón),42 or fossilized as topographical names (Tedeja),43 and
sometimes even the place-name changed, as with the transformation of Buradón
into Bilibio. All are different signs of a process of abandonment, demonstrated
archaeologically in Tedeja and Buradón. Their decadence would be linked to the
formation of a new social system, which implied a change in the reference-points
of power. Feudal lords exercised their dominion through the control over the
peasant working process and not over the socio-political resources of communi-
ties, and through the participation in a new central political apparatus organized
around feudal relationships.
There were other hilltop sites of smaller size that controlled minor areas. The
recognition of these territories as political districts named alfoces by the Castilian
39
It was the seat of a royal agent, with some fiscal and judicial functions, in the middle of the
eleventh century. Colección Diplomática San Salvador de Oña, ed. Álamo, doc. 27.
40
Lecanda Esteban, ‘Mijangos’.
41
R . Bohigas Roldán and others, ‘Tedeja y el Control político del territorio del Norte Burgalés
en Época Tardorromana, Visigoda, Alto y Plenomedieval’, in V Congreso de Arqueología Medieval
Española, 2 vols (Valladolid, 2001), I, 49–56 (pp.54–55).
42
See Martín Viso, ‘Poder político’, pp.550–51.
43
The last mention linked to a royal agent in Tedeja is dated in 1127; Los Cartularios Gótico
y Galicano, ed. Ruiz de Loizaga, Gótico, doc. 149.
Page 196
Ordernr. 061297 P1
counts means that they were considered as cells of the territorial articulation of the
new authority. However, Castilian central power did not possess the means to create
these districts, so its aim was the control of some processes and areas in order to insert
them in its network.44 This integration did not mean that all situations were similar,
or that all the districts developed an identity as alfoz; nor did it mean the general
implementation of this pattern. In the areas where supralocal units were stronger,
it was generally expedient to include them into the new political construction.45
The hilltop sites could function as one of the main types of central place on a
small scale, but in other areas monasteries could carry out this function.46
It is necessary to analyse how the alfoces located in districts that initially did
not constitute the axis of the formation of Castilian power worked. One of them
is Siero, quoted as an alfoz in the tenth and eleventh centuries, including a great
part of the valley of Rudrón into its territory. The latter area probably lay outside
its initial area of activity, so there some small hilltop sites, like Castrillo de Butrón
and probably Sedano, were inserted into it. Another very similar case is Alfania,
controlling the eastern part of Valderredible — an area without hilltop sites —
during the reign of García III (1035–54), when this zone was a frontier area.47
This expansion came together with a change of the place-name in favour of
Paredes Rubias,48 a modification linked to the construction of the small fortress
near the village of Báscones de Ebro.49 Both examples show how hilltop sites and
their territories were bases for the formation of new districts during the imple-
mentation of Castilian authority. This involved a new arrangement of the hilltop
44
This hypothesis can be deduced through some recent studies, such as Álvarez Borge,
‘Estructuras’; Escalona Monge, Sociedad y territorio; or I. Martín Viso ‘Pervivencia’.
45
See C. Estepa Díez, ‘El alfoz castellano en los siglos IX al XII’, in En la España Medieval, IV:
Estudios Dedicados al Profesor D. Ángel Ferrari Núñez, 2 vols (Madrid, 1984), II, 305–41; Álvarez
Borge, Monarquía feudal, pp.7–54; Escalona Monge, Sociedad y territorio, pp.85–184.
46
A good example is the monastery of Valpuesta. The role of the ecclesiastical centres is well
known; see E. Peña Bocos, ‘Ecclesia y Monasterium: elementos de ordenación de la sociedad de la
Castilla altomedieval’, in Señorío y Feudalismo en la Península Ibérica (siglos XII– XIX ), ed. E. Sarasa
Sánchez and E. Serrano Martín, 4 vols (Zaragoza, 1993), III, 379–98.
47
Colección Diplomática San Salvador de Oña, ed. Álamo, doc. 35 (1048.04.10).
48
Although the term had appeared in the foundation endowment of San Salvador de Oña,
dated 1011, it was more common from the second half of eleventh century. Burgos, doc. 34.
49
R . Bohigas Roldán, ‘Fuentes arqueológicas y organización social del espacio en el reino de
Castilla’, in Del Cantábrico al Duero, ed. García de Cortázar, pp.75–121 (p.115).
Page 197
Ordernr. 061297 P1
sites, with fission and hierarchization promoted by central power.50 The previous
balance between residential and defensive areas was broken, and the sites were
essentially transformed into fortresses. It is very difficult to detect the defensive
structures, because they were probably built of timber, whose remains have not
been recognized by archaeologists. The local leaders that exercised domination
over communities in the eighth and ninth centuries probably did not develop
complex means of extracting large surpluses, so they had to resign themselves to
using these simple types of construction. The eleventh- and twelfth-century
charters identify some of these old hilltop sites with castles, showing what their
main function was in the eyes of the clerks,51 although stone structures were never
erected in many of these sites.
At the same time, a gradual abandonment of many hilltop sites can be traced,
with an intermediate phase of use as religious centres, as shown by the excavations
carried out at Castros de Lastra (Caranca).52 Different factors can explain this trend.
Once the aristocratic groups were transformed into feudal lords, they laid aside the
control of hilltop sites, which had been the centres for some of the socio-political
roles of communities. Social power was based on the domination of the peasant
working process, the collection of feudal dues and political service to the Castilian
king. The effective exercise of the royal power was focussed on some fortresses, such
as Término, Tedeja, Piedralada or Aguilar, where an agent of the king was located,
the so-called dominante, normally an aristocrat, with military and judiciary func-
tions. In addition, the decay of the hilltop settlements must be connected with the
gradual process of settlement nucleation, which had, as its main effect, the
generalization of the village as a settlement form throughout the region.
Although this trend was general, there are some cases of hilltop sites, such as
Ruanales, Medina de Pomar, Moradillo del Castillo or Villalba de Losa, which
have survived as villages to the present day. The fact that present-day nuclei lie
over the previous sites makes accurate understanding of their medieval phases
difficult, but their location forces us to consider them as hilltop settlements — a
situation that high medieval sites normally avoid — and some were central places
of alfoces (Ruanales and Moradillo del Castillo). Be that as it may, early defensive
structures disappeared or were replaced in favour of stone walls, as at Medina and
50
See Escalona, this volume.
51
It was the case of Berbeia; Los Cartularios Gótico y Galicano, ed. Ruiz de Loizaga, Gótico,
doc. 136.
52
F. Sáenz de Urturi, ‘Castros de Lastra: XX campaña de excavaciones’, Arkeoikuska (1993),
53–62.
Page 198
Ordernr. 061297 P1
Villalba, both royal boroughs from the twelfth century. Their survival as simple
villages was, however, more common, although the site’s survival was sometimes
linked with the continuation of a supralocal unit based on the hilltop site until
the thirteenth century, above all in those areas removed from the great foci of
lordship.
Probably the last reason explains why the hilltop continued to be an important
site in the mountain areas. The communities of the southern European highlands
have been characterized as very strong and cohesive, allowing the conservation of
a settlement pattern that had been abandoned in other areas.53 Pernía — a moun-
tain district located in the high valley of the river Pisuerga — may stand as a good
example to illustrate this situation in northern Castile. In spite of the problems
of the archaeological data and the scant early medieval written references, three
territories focussed on hilltop sites seem to be revealed: Piedras Negras,54 Resoba55
and Tremaya.56 There were probably another two: Castillería — whose name
derives from the word for castle — and Mudá, mentioned as an alfoz in the
delimitation of the bishopric of Palencia in 1059.57 All of them controlled narrow
stretches of valley, where complementary use of lowlands and uplands can easily
be developed. The survival of pre-Roman place-names of some of the central
places and villages seems to be connected with a pre-medieval origin. The absence
of geo-strategic interests, because there was no important route crossing the dis-
trict and Pernía was never an area whose control was contested between different
53
About mountain areas of the Iberian north-west, see M. Fernández Mier, Génesis del Terri-
torio en la Edad Media: Arqueología del Paisaje y Evolución Histórica en la Montaña Asturiana
(Oviedo, 1999) and J. A. Gutiérrez González, ‘Sobre los orígenes de la sociedad asturleonesa:
aportaciones desde la arqueología del territorio’, Studia Historica: Historia Medieval, 16 (1998),
173–97.
54
J. Pérez de Urbel, Historia del Condado de Castilla, 3 vols (Madrid, 1945), III, doc. 129bis;
J. L. Martín, Orígenes de la Orden Militar de Santiago (1170–1195) (Barcelona, 1974), doc. 113.
It could be identified with the hilltop site of Peñas Negras, near the village of Arbejal.
55
Pérez de Urbel, Historia, III, doc. 129bis; Documentación de la catedral de Palencia
(1035–1249), ed. T. Abajo Martín (Palencia, 1986), docs 82 and 91; Martín, Orígenes, doc. 118.
56
Colección Documental del Archivo de la Catedral de León, IV (1032–1109), ed. J. M. Ruiz
Asencio (León, 1990), doc. 952, and Martín, Orígenes, doc. 183. The place was probably near
Santibáñez de Resoba.
57
Documentación de la catedral de Palencia, ed. Abajo Marín, doc. 9. See also Colección
Diplomática del Monasterio de Sahagún, II (1000–1073), ed. M. Herrero de la Fuente (León,
1988), doc. 530. Some historians have thought that there was a castle inside the territory, although
its location remains unknown even now.
Page 199
Ordernr. 061297 P1
polities, supports the hypothesis of a communal origin. That idea was reinforced
thanks to the presence of a homogeneous royal lordship (realengo*) up to the end
of the twelfth century, whereas the local aristocratic elite did not have feudal
status. The kings claimed a general dominion over the communities58 and con-
trolled the activity of some local monasteries under their patronage.59 The process
of feudalization came together with a trend to develop commercial, extensive
stock raising in these areas from the thirteenth century. The definitive break in
territorial organization must be dated to around 1200, implying an enduring
tradition compared with other areas of northern Castile (Fig. 8.5).
This territorial pattern, based on strongly cohesive communities, also appears
in other highland areas of northern Castile (Sotoscueva, Valdebodres, Espinosa)
and in different areas of the Iberian peninsula. The weak hierarchy among the
different hilltop sites, the stability of territories and the absence of a powerful
indigenous class of feudal lords explain the highlands’ experience. The key might
be found in the communities’ adaptation to a difficult environment. They chose
control of some territories organized around stretches of valley and probably
based on seasonal small-scale transhumance and stock raising. This pattern did
not need great initial investment but did require the maintenance of strongly
cohesive communities to allow the pastures’ use by all members and their defence
against other groups. This model was very stable because any change could only
be made by the integration of other similar systems or by inclusion in a different
system. The first possibility was very difficult, because of the inability of local
elites to transcend the limits of the communal framework. Thus the status of the
elites was connected with the exercise of prestigious social functions and not with
the exploitation of large estates, which would require the acquisition of new
pasturelands in the hands of other communities.
Finally, although the occupation of hilltop sites was a general phenomenon in
northern Castile, it is necessary to emphasize their heterogeneity, which can be
summarized as falling into three different patterns: the defensive point created by
the communities and associated with a nearby settlement; the defensive area
placed on the summit, without any settlement, articulating a valley community
in the highlands; and the complex defensive structures begun by a central political
58
The hilltop sites were identified as royal castles and some taxes would be focussed on them.
Documentación de la catedral de Palencia, ed. Abajo Martin, docs 48 and 82.
59
For instance, San Miguel de Infravillas, San Salvador del Monte, San Salvador de
Cantamuda and Santa María de Lebanza; Documentación de la catedral de Palencia, ed. Abajo
Martín, docs 14, 82 and 91.
Page 200
Ordernr. 061297 P1
Figure 8.5. Hilltop sites and territories in the district of Pernía.
apparatus, with its residential functions located on the plain. The first two possi-
bilities were the most usual and they implied an initiative by communities. The
occupation of the hilltop sites must be connected with the transformation of the
social basis of aristocratic groups after the collapse of the Roman system. These
changes could take different forms, according to previous regional conditions, and
the occupation of hilltop sites in varying forms was their consequence.
Page 201
Ordernr. 061297 P1
Page 202
Ordernr. 061297 P1
T HE E NDING OF THE R OMAN C ITY : THE C ASE

OF C LUNIA IN THE N ORTHERN P LATEAU OF S PAIN
Adela Cepas
S
everal essays in this book deal with areas with a Roman past, areas where
early medieval communities emerged from a complex background shaped
by the dismantling of the social and political framework provided by Rome
and the appearance of new political identities and forms of organization.
Investigating the obscure origins of early medieval communities through both the
archaeology and the written evidence is a hard but crucial task, as most authors in
this volume remark.
In this essay my aim is to suggest that a deeper examination of the relationship
between the Roman political framework and the local communities living under
it may provide rich insights into the kind of processes from which their early
medieval successors took shape. Drawing on the case of Clunia — the Roman
capital of the conventus iuridicus* Cluniensis in the Spanish northern plateau —
I will explore changing political identities and notions of community across the
Imperial period with an emphasis on how, and to what extent, the notion of
Roman civitas* moulded political identities in Rome’s heyday, and how much the
ending of this model in the late Roman and sub-Roman periods reveals about
alternative social and political aggregations, about, in other words, the melting pot
of early medieval communities.
Political Identities Within the Roman Empire: A Fragile World?

In the Roman Empire ethnic matters were emphatically separated from politics, to the
extent that they became completely independent. An individual could keep up everything
pertaining to his native people and culture and only be Roman in so far as he was a
political being, a citizen. Society was based on civil law, which was seen as an agreement
Page 203
Ordernr. 061297 P1
188 Adela Cepas
between members of society aimed at achieving the common good. And this idea of the
common good, the res publica, came to be considered as being something higher, more
important than the ethnic community.1
It is worth quoting this lengthy extract since it makes a useful introduction to the
plurality of the Roman Empire. The many different peoples at different stages of
cultural and political development, the complexity of social and economic rela-
tionships, the different religious faiths, and so on, all gathered together into an
immense political entity, are a fine example of a multi-ethnic, multicultural and
multiconfessional state, under the solid unifying umbrella of Romanitas.
There is a relatively widespread opinion that the decadence of the Roman
Empire brought with it a return of certain cultural traditions that had existed in
the western Roman Empire before the conquest and lived on through the follow-
ing centuries. Manufacture of certain objects (mainly ceramics, sculpture, metal-
work) in pre-Roman style and form seems to have taken place from as early as the
second half of the second century, when Rome’s political and economic power was
at its height. The so-called ‘Celtic survival’ is not just a feature of Europe’s far
north. In Spain, pieces such as the slabs set up by the Vadinian community have
also been seen as examples of this Roman ‘fragility’. These are a set of second- to
third-century funerary and votive inscriptions found on both sides of the Can-
tabrian mountains, across the boundaries of the modern provinces of Asturias and
León; neither their form, nor the script, nor the personal names in them fit into
the classical format.2 No wonder that they have been used as examples of how
little Romanization really permeated certain indigenous communities.
The three main reasons usually forwarded for the ‘Celtic survival’ — stimula-
tion by contact with other less Romanized people, economic decline and the
increase of nationalism — were all rejected by MacMullen, according to whom
this happened just because ‘the work of conquerors, remarkable as it certainly was,
suffered from a kind of fragility and superficiality which should not be forgotten’.3
Certainly, the processes running in the highest spheres need to be counterbal-
anced with the more varied local realities. It is indeed known that Rome’s success
stemmed from its ability to govern a vast territory with a relatively elementary
1
G. Pereira Menaut, ‘Formación técnica vs. humanismo: Aproximación crítica’, Revista
Mientras Tanto, 68–69 (1997), 135–48 (pp.144–45).
2
Mª. C. González and J. Santos, ‘La epigrafía del conventus Cluniense, I: Las estelas
vadinienses’, Memorias de Historia Antigua, 6 (1984), 85–111.
3
R . MacMullen, ‘The Celtic Renaissance’, Historia, 14 (1965), 93–104 (p.104).
Page 204
Ordernr. 061297 P1
THE ENDING OF THE ROMAN CITY 189
administrative establishment.4 The army — which not only conquered but was
also directly involved in tasks more properly belonging to the work of pacification,
administration and exploitation of the territory — together with a small number
of civil servants, organized into various ranks depending on the tasks they had to
accomplish, were the pillars on which Rome leant to manage its huge empire.5 But
this was not enough to govern and economically exploit a territory and to make
a large population pay its taxes. It is agreed that the very key of Roman territorial
control was a system of cities, which governed themselves, financed themselves
and, in their turn, financed Rome. It is in the cities (civitates) that the local
population became directly involved.
A city’s population comprised both the people living in the town (urbs) and
those dispersed across the territory (ager). The town was not only a political and
religious centre where the local curia* lived and met, but also a market and a
stronghold of defence. Apart from the urban core, the territory hosted a popula-
tion of varying legal and social status, scattered in settlements which also varied
in terms of type, functionality and position in relation to others; some were of
more urban character, like vici*, pagi, fora (marketplaces) or mansiones (road
stations), but in the rural areas there was a veritable explosion of new settlements
connected with the exploitation of the territory such as villae* (agricultural
estates), mining sites and saltpans, pottery workshops and factories. This settle-
ment network was reinforced by a dense pattern of roads connecting the main
urban centres in each province that guaranteed political and fiscal control of the
conquered population. 6 We need to look into the heart of civitates in order to
identify those communities that, by being included within the city structures,
became integrated in the new political order established by Rome.
4
Roman administration may seem elementary by modern standards, but it was highly
sophisticated if compared with that of its predecessors and successors.
5
F. Millar, The Roman Empire and its Neighbours (London, 1967); F. Millar, The Emperor
in the Roman World, 31 BC – AD 337 (London, 1977); K. Hopkins, Conquerors and Slaves
(Cambridge, 1978); H. G. Pflaum, Les carrières procuratoriennes équestres sous le Haut Empire
romain, 3 vols (Paris, 1961).
6
P. López Paz, La Ciudad Romana Ideal, vol. I.1: El territorio (Santiago de Compostela,
1994); A. Rey Losada, La Ciudad Romana Ideal, vol. II: El grupo poblacional (Santiago de
Compostela, 2003); M. I. Finley, The Ancient Economy, 2nd edn (London, 1985), pp.177–207;
P. Garnsey and R . Saller, The Roman Empire: Economy, Society, Culture (London, 1987).
Page 205
Ordernr. 061297 P1
190 Adela Cepas
Administrative Frameworks and the Territorial Role of Cities in

Northern Hispania
Following the Roman conquest Hispania was organized in a three-layered admin-

istrative system. The peninsula was divided into three provinces: Baetica, Lusita-
nia and Citerior. The territory of each province was then divided in legal districts
(conventus), comprising a number of civitates; the latter were the units into which
local communities were integrated, peacefully so in many cases — since in all
probability most communities continued to live in their previous homes — but with
a measure of violence wherever Rome forced the population to move from their
old settlements in the hills to new ones on the plain, closer to roads and cities.
This system was rearranged and simplified by Diocletian in the late third century.
The Iberian Peninsula became part of the diocesis Hispaniarum, to which some
North African territories were added. Seven provinces were created — Baetica,
Lusitania, Carthaginensis, Callaecia, Tarraconensis, and the African Mauritania
Tingitana7– mainly by splitting into three the huge Citerior (Fig. 9.1). The interme-
diate administrative units — the conventus probably — disappeared. These changes
mostly affected the upper levels but not the internal workings of the civitates.
The fourth century was a haven of peace between the troubled third (especially
in its second half) and the even more disturbed fifth century, characterized by the
confrontation between the usurper Constantine III and the official emperor
Honorius, the intrusion of Sueves, Vandals and Alans in 409, and the Goths’
increasing interference in the Iberian peninsula. When Roman rule faded out and
the provincial government and the army disappeared, a varied range of cities
remained that were left to survive as best they could on their own. Even if Roman
administrative structures largely continued throughout the fifth century, as it has
been recently argued,8 urban centres and their territories were in effect the main
administrative framework for much of northern Hispania.
Between the first and third centuries AD, Clunia was the capital of one of the
intermediate administrative districts in the province Hispania Citerior, the
conventus iuridicus Cluniensis. This was a large district stretching from the
7
The limits of Baetica and Lusitania remained practically unchanged. In the rest, the new
provincial boundaries obliterated those of the former conventus. Moreover, epigraphic mentions
of conventus and their administrative offices disappear in the fourth century. J. Arce, El último siglo
de la España Romana (274–409) (Madrid, 1986), pp.31–41.
8
J. Arce, ‘Los gobernadores de la Diocesis Hispaniarum (ss. IV – V d.C.) y la continuidad de las
estructuras administrativas romanas en la Península Ibérica’, Antiquité Tardive, 7 (1999), 73–83.
Page 206
Ordernr. 061297 P1
Figure 9.1. Roads, towns and late provincial divisions in Roman Spain.
eastern side of the modern province of Asturias to Bilbao, and from the Cantabric
shore to the Spanish Central mountains. It engulfed a great number of native
peoples and settlements of very diverse cultural traditions (Fig. 9.1).
Why Clunia was chosen as the conventus capital is far from being fully ex-
plained. Although there surely was a native settlement in the vicinity — not yet
detected — Roman Clunia was newly founded on top of a prominent 1023 metre-
high plateau, in a location relatively detached from the main road network. It was
on no communication network. A detour from the main road linking Asturica
(Astorga) and Bracara (Braga) with Tarraco (Tarragona) via Caesaraugusta (Zara-
goza) was built to fit Clunia into this pattern. Likewise the city’s connections with
other major urban centres had to be established thereafter by means of a dense
web of secondary roads.9
Clunia’s archaeological story is a familiar one: a fast urban development in the
early empire followed by a late Roman period marked by the change of function
9
J. A. Abásolo Álvarez, Las vías romanas de Clunia (Burgos, 1978).
Page 207
Ordernr. 061297 P1
192 Adela Cepas
and/or abandonment of most of its structures. Like most major towns in the
Duero plateau,10 Clunia flourished very rapidly in the first century AD. Its best-
known public facilities are a theatre, forum, urban street-plan, two large public
baths, and a substantial building of uncertain function near the forum as well as
several stately houses (domus). Urban development continued at a dynamic pace
in the following centuries, with a series of reformations both in the forum and in
other areas of the city. By the late empire, though, most of the public buildings
had lost their former function. The abundant material culture unearthed by Palol
since 1958 is in sharp contrast with the use people made of the city’s buildings.
Glass and bronze objects, coin hoards spanning the mid-third to fifth centuries,
and a mass of pottery have been recovered. The especially abundant Late Hispanic
Red Slip Ware is found in the later phases of houses and other buildings, often
marking a reuse, rather than normal continued occupation. This happens indis-
tinctively in one of the forum temples, in house number 3, in the small baths by
the forum and in the large ones called Los Arcos, where findings of pottery
moulds suggest the facilities had been turned into a pottery workshop. Likewise,
a fifth- to seventh-century cemetery invaded the urban area east of the forum.
After the fifth century Clunia, to be sure, did not look much like the standard
Roman city it had once been.11
Archaeological material from Clunia spans the whole Roman period with no
clear sign of widespread abandonment, but there certainly was a profound change
of character between the early and late empire, expressed, among other factors, by
a change of attitude in the local elites, who dropped public building construction
and maintenance. Similar processes in other towns of central Spain like Valeria,
Segobriga, Ercavica, or Vxama have usually been explained by reference to the
third-century political and economic crisis.12 In Clunia’s case, the ongoing process
surely reached a turning point with the late third-century suppression of the
conventus division, which removed its capital character and downgraded it to the
status of a regular city within the new province.13
10
A. Cepas Palanca, Crisis y continuidad en la Hispania del siglo III, Anejos de Archivo
Español de Arqueología, 17 (Madrid, 1997), pp.181–86; R . F. J. Jones, ‘A False Start? The Roman
Urbanization of Western Europe’, World Archaeology, 19 (1987), 47–57.
11
P. de Palol, ‘Noves dades arqueologiques sobre els darrers segles de Clunia’, orig. pub. 1978;
repr. in P. de Palol and others, Clunia 0: Studia Varia Cluniensia (Burgos, 1991), pp.295–300.
12
Cepas Palanca, Crisis y continuidad, pp.230–31.
13
P. de Palol, Clunia: Historia de la ciudad y guía de las excavaciones, 6th edn (Burgos, 1994),
p.140.
Page 208
Ordernr. 061297 P1
Clunia’s later phases are poorly understood yet. A Christian community has
been hypothesized after a glass bowl with Christian symbols was found back in the
1960s in the reused bath at Los Arcos.14 This may be connected to the aforesaid
sub-Roman cemetery, but no buildings of distinctively religious typology have
been detected. Clunia seems to have enjoyed no prominent position in the
Visigothic period: no coins were minted there; no textual reference has been
preserved. Moreover, it failed to become an episcopal see, which was set in nearby
Vxama Argaela, as a confirmation of Clunia’s loss of status in the late Roman
period. There is not much to be said about the early medieval phase. A population
may have remained, whether on the city’s platform or in the lowlands around it.
If so, their location shifted to a lower promontory in the south-west, modern
Coruña del Conde. This new location retained the ancient name and a certain
notion of territorial centrality. In the tenth century it became the capital of the
largest district in the County of Castile and suffered several attacks by the
Cordoban Muslim armies.15
This sketchy account of Clunia’s trajectory should be enough to argue that the
striking changes the city underwent across antiquity must have affected its rela-
tionships with the territory it controlled and the population living there. Yet
community membership and political identities are hard to investigate because of
the rock-solid message of city allegiance that stems from the textual records, which
nearly always originate in the centre. It is through the interstices of this apparently
uncontested discourse of political unanimity that we must proceed.
An Epigraphic Exploration of Political Identities in the Clunia Area
The local populations of Roman times have left their traces in a multitude of
mostly private inscriptions which provide us with a host of data such as deity
names, kin-groupings, or personal names, as well as a whole web of family relation-
ships left to us mainly on funerary inscriptions. This material is our best available
means of getting closer to local political identities.
A massive literature has analysed the connections between the latinization of
personal names and changes in the legal status of native individuals and communi-
ties, from their initial situation of foreigners (peregrini) within the Roman state
14
P. de Palol, ‘Un vidrio tallado, con temas cristianos, de Clunia’, in de Palol and others,
Clunia 0, pp.347–54.
15
J. Escalona, ‘Comunidades, territorios y poder condal en la Castilla del Duero en el siglo X ’,
Studia Historica: Historia Medieval, 18–19 (2000–01), 85–120, and his essay in this volume.
Page 209
Ordernr. 061297 P1
194 Adela Cepas
to personal or collective promotion. Those who achieved Roman citizenship

tended to restyle their native names into a more appropriate, Roman-like format.
An early inscription with Latin names probably indicates an individual from
outside Hispania whereas, later, more and more natives bear Latin names so the
distinction becomes blurred.16 Most importantly, members of the local elites
reinforced their local leadership and strove to join other higher social and political
spheres by becoming part of the administration of a city, a juridical district, a
province, the army or the Roman state.
The lowest step in this ladder of promotion was the city, but this was far from
being a simple environment. Unfortunately, there is no direct evidence we can use
to plot the boundaries of first- to third-century Roman municipia* in these areas.
Yet the distribution of privileged towns suffices to show that they surely had very
large hinterlands. When compared with the regional network of late pre-Roman
Iron Age oppida*, the Roman pattern clearly means that only some of them,
together with a limited number of new foundations, made it to urban status and
developed proper territories. Those communities that failed either disappeared
or became lesser status sites (vici) embedded in an urban ager. The Roman city,
therefore, was the centre of a complex system of overlapping ethnic and political
identities — kin-groups, urban membership, Roman citizenship — which can be
observed in the inscriptions. Let us look at these in detail.
As expressed on inscriptions, native personal names normally feature as a first
name plus another in the genitive, recording filiation. Not infrequently a third
element is added, often in an adjectival form, more rarely in the genitive plural,17
which is taken to express belonging to some sort of grouping, whose character is
16
As regards chronology, we should bear in mind that we are dealing here with the northern
part of the Iberian Peninsula where an early inscription usually dates from the change of era (i.e.
before and after 1 AD ). For the transition from indigenous names to Latin ones, see G. Pereira
Menaut and J. Santos Yanguas, ‘Sobre la romanización del noroeste de la península Ibérica: las
inscripciones con mención de origo personal’, Revista de Guimaraes (1980), 120–22. For the
difficulties associated with learning about people’s juridical status through epigraphy, see
G. Alföldy, ‘Notes sur le droit de cité et la nomenclature dans l’Empire Romain’, Latomus, 25
(1966), 37–57.
17
Names ending in the genitive plural refer to a group of people who were related through
their father and mother, as documented in the Indo-European area of Spain on Celtiberian
tesserae, and in inscriptions of the type Aecus Aploniocum Lougi f(ilius) Cluniensis, Aploniocum
being the name ending in genitive plural. The inscription was found in Lusitania (Alconetar,
Cáceres); see M. E. Ramírez Sánchez, Epigrafía y organización social en la region celtibérica: los
grupos de parentesco (Las Palmas de Gran Canaria, 2001), p.413, A.91.
Page 210
Ordernr. 061297 P1
much debated.18 Terms such as ‘tribal organizations’ and ‘suprafamily organiza-

tions’, as well as names ending in the genitive plural, are very widely used in
Spanish historiography to describe native social groupings in the Celtic-speaking
areas of Iberia as well as in large zones of the north and north-west of the Iberian
Peninsula. Native ‘peoples’ — in the broadest and loosest sense of the word — are
seen as nested aggregates of smaller, self-defined groupings, normally expressed as
gentes, which in turn may engulf other gentes.
Now, these terms probably refer to kin relationships (cognationes*). Their
physical dimension is uncertain. A relationship has been put forward, mainly for
the Galician far north-west, between names in the genitive and castella* (hill
forts) to suggest that kin-groups could also be residential units.19 Yet kin networks
may be perfectly operative social structures without a residential character. They
may constitute forms of non-territorial community bonds, spreading across
several settlements and territories, to which some of their members belong as part
of the local population. A flexible approach seems preferable after growing
criticism eroded the established image of native pre-Roman societies as ‘tribal’,
egalitarian, kin-based organizations, in favour of a more sophisticated combina-
tion of kin-based links with a greater political and economic complexity, and some
degree of class differentiation.20
Clunia lay in the area of the Arevaci, a pre-Roman people belonging to Celti-
beria.21 The (mostly funerary) Latin inscriptions from Clunia, dated to the first
century AD or, at the very latest, the beginning of the second record people who
expressed ascription to a group of the kind expressed by native names in the
genitive plural, although the former amount to just five references out of a total
18
M. C. González Rodríguez, Las unidades organizativas indígenas en el área indoeuropea de
la Península Ibérica (Vitoria, 1986); F. Beltrán Lloris, ‘Un espejismo historiográfico: Las
organizaciones gentilicias hispanas’, in I Congreso Peninsular de Historia Antigua (Santiago de
Compostela, 1986), ed. G. Pereira Menaut, 3 vols (Santiago de Compostela, 1988), II, 197–237
(pp.218–29); M. Salinas de Frías, ‘Sobre las formas de propiedad comunal de la cuenca del Duero
en época prerromana’, Veleia, 6 (1989), 103–10; J. Santos Yanguas, ‘1985–1994: Un decenio
fructífero en la investigación de las estructuras sociales indígenas del área indoeuropea’, Veleia, 12
(1995), 125–49.
19
M. L. Albertos Firmat, ‘Organizaciones suprafamiliares en la Hispania Antigua’, Studia
Archaeologica, 37 (1975), 5–66.
20
F. Burillo Mozota, Los Celtíberos: Etnias y estados (Barcelona, 1998).
21
Pliny, Natural History, III.27, in Naturalis Historia, ed. C. Mayhoff (Leipzig, 1906): Clunia
Celtiberiae finis; J. Mª. Gómez Fraile, ‘Celtiberia en las fuentes grecolatinas: Replanteamiento
conceptual de un paradigma obsoleto’, Polis, 8 (1996), 143–206.
Page 211
Ordernr. 061297 P1
196 Adela Cepas
220 pieces.22 Even individuals who enjoyed Roman citizenship sometimes re-
corded this ascription, with no apparent interference between this private display
and the ius civile,23 as in the case of L. Valerius C.f. Gal(eria) Crescens Bundali-
co(rum).24 Similar data can be found in the nearby town of Nova Augusta (Lara
de los Infantes),25 whose large epigraphic collection seems to span from the late
first century to the middle of the third, with a particularly splendid period
between 130 and 200 AD.26
So it seems that for local people, who were buried in Clunia or its immediate
vicinity, it was fit to record that they belonged to one of those groupings, whatever
their character. Now, when it comes to political identities, what they all had in
common was their belonging to a civitas. This becomes apparent when we can see
people acting away from home. Local citizenship imposed itself on the collective
mentality of the native peoples to such an extent that when physically outside the
civitas territory, their belonging to a particular native grouping ceased to be a valid
designation. For example, a detailed analysis of inscriptions from Asturias,27 in
northern Spain, shows that if an individual died, or dedicated a votive inscription,
within his civitas territory, only the group-membership was expressed. If, on the
other hand, he died outside the civitas territory, both the grouping and the origo
22
Ramírez Sánchez, Epigrafía y organización social, p.327, A.32: Aeggu(m); p.325, A.33:
Bundalico(rum); p.327, A.34: Morcicum Aquilliorum; p.328, A.36: Usseiticu(m), always in a
funerary context.
23
G. Pereira Menaut, ‘Cognatio Magilancum: Una forma de organización indígena de la
Hispania indoeuropea’, in Actas del V Coloquio sobre Lenguas y Culturas Prerromanas de la
Península Ibérica, ed. J. Untermann and F. Villar (Salamanca, 1993), pp.411–24 (pp.423–24).
24
‘Lucius Valerius Crecens son of Gaius of the voting tribe Galeria of the Bundalico(rum)’,
with the same Celtiberic ending in genitive plural but now in Latin: Ramírez Sánchez, Epigrafía
y organización social, p.325, A.33; P. de Palol and J. Vilella, Clunia II: La Epigrafía de Clunia
(Madrid, 1987), no. 217.
25
H. Gimeno and M. Mayer, ‘Una propuesta de identificación epigráfica: Lara de los Infantes
/ Nova Augusta’, Chiron, 23 (1993), 313–21; Ramírez Sánchez, Epigrafía y organización social,
p.327, A.26: Alticon; p.318, A.27: Belvicon; p.319, A.28: Cabuecon; p.321, A.29: Caelaon; p.322,
A.30: Elaesic(um); p.324, A.31: Moenic(u)m.
26
J. A. Abásolo Álvarez, ‘Las estelas decoradas de la región de Lara de Los Infantes: Estudio
Iconográfico’, Boletín del Seminario de Arte y Arqueología de la Universidad de Valladolid, 43
(1977), 61–90.
27
G. Pereira Menaut, and J. Santos Yanguas, ‘Ensayo de sistematización de la epigrafía romana
de Asturias: Las unidades gentilicias’, Boletín del Instituto de Estudios Asturianos, 105–106 (1982),
87–110.
Page 212
Ordernr. 061297 P1
(civitas of provenance) were recorded. Moreover, for Roman citizens who died
away from their homeland, only the civitas of origin was noted down; no native
grouping was needed.
Likewise, in the inscriptions from Clunia, references to Clunienses as Clunian
citizens only occur when related to officials of the provincial or colonial ad-
ministration, as in a fragment of a bronze plaque containing two lines of a first-
century AD tabula patronatus* that reads [— Cluniens]es ex H(ispania)
C(iteriore).28 Also, a tabula patronatus which has been preserved whole records
that in 40 AD the Clunienses ex Hispania Citeriore made an agreement of hospi-
tality (hospitium*) with C(aius) Terentius Bassus C(aii) f(ilius) Fab(ia) Mefanas
Etruscus praefectus ala Augusta together with his sons and heirs, formalized by two
Clunian delegates.29 And in 222 AD, in a tabula patronatus in bronze, found in
Rome in 1776, the Concilium Conventus Cluniensis adopted G(aius) M(arius)
P(udens) Cornelianus leg(atus) leg(ionis) c(larissimus) v(ir) as patron (patro-
nus*).30 Private documents with a public dimension also mention the civitas, as
the dedication of a servus* (slave) of the res publica (city) of Clunia to the numen
(deity) of the city theatre, found on the site of the theatre, or the fragments of a
marble plaque in which someone, probably a freedman of Hadrian’s, honours the
T(utelae) Colon(iae) Cluniensium.31
28
‘Clunienses that form part of the Province of Hispania Citerior’. This fragment was found
in the forum, in the same sector where two other similar fragments appeared (nos 113 and 114),
possibly all of them from a single bronze plaque on which, according to the publishers, Clunia’s
municipal laws may have been inscribed; de Palol and Vilella, Clunia II, no. 115.
29
‘Giaus Terentius Bassus Mefanas Etruscus son of Gaius of the voting-tribe Fabia, prefect
of the ala Augusta’: de Palol and Vilella, Clunia II, no. 116. A. D’Ors, Epigrafía Jurídica de la
España romana (Madrid, 1953), p.373, no. 21, interprets this as meaning that the agreement was
made between the Clunians and a specific individual and his heirs. The tablet, which is well
preserved, measures 38 x 28 x 0.5 cm and was found by a ploughman in the ruins of Clunia in
1887. The tablet has a hole in each of its four corners so that it could have been hung on a wall,
probably within the precincts of the forum. The Ala Augusta was stationed in the Iberian
Peninsula; cf. P. Le Roux, L’armée romaine et l’organisation des provinces ibériques d’Auguste à
l’invasion de 409 (Paris, 1982), pp.90–91.
30
‘The Council of the Conventus Cluniensis adopts Gaius Marius Pudens Cornelianus
commander of a legion, clarissimus vir, as patron’: de Palol and Vilella, Clunia II, no. 117; CIL,
VI, 1454; D’Ors, Epigrafía Jurídica, p.377, no. 27, suggests that the tablet was found in Rome
because the patron concerned probably lived there.
31
‘The tutelary goddess of the colonia of Clunians’: de Palol and Vilella, Clunia II, nos 21 and
22 respectively. Inscription no. 22 came to light in a building along with columns, perhaps a
temple dedicated to Isis, since a statue of this goddess appeared next to this inscription.
Page 213
Ordernr. 061297 P1
198 Adela Cepas
By contrast, no private individuals styled themselves Clunians when having a

funerary or votive inscription put up within the city territory. Those who did
express their origo were foreigners who died away from their homeland. Thus,
Aius Antonius dedicated a funerary inscription to his wife Attua Boutia daughter
of Boutius from Intercatia who died in Clunia, far away from her own city,
Intercatia.32 Similar cases include a person from Vindeleia,33 or a freed woman
from Caesaraugusta,34 or two veterans from Emerita and Caesaraugusta.35 In an
altar dedicated to the Matres Brigeaecis by L. Aelius Phainus, the city name fea-
tures as an adjective to the deity’s name,36 which surely also applies in the case of
a similar dedication to the Matres Endeiteris.37 Conversely, when dying or wor-
shipping their gods away from their place of origin, the Clunians represented
themselves as members of their civitas, and they expressed their origo accordingly.
This is particularly clear in a somewhat homogeneous group of funerary slabs
found in Vicus Spacorum (modern Vigo, in south-western Galicia) and dated to
the second century AD onwards.38 Text-structure, iconography and ornamenta-
tion all point to people coming from outside Roman Callaecia, or the conventus
Lucensis, to which Vicus Spacorum belonged. Three members of a family can be
identified who state their origo as Clunian. Aurelia Materna dedicated a slab to
her husband Quintus Arrius Mansuetus son of Geneus from Clunia who died in
Vicus Spacorum aged forty (no. 39). The dead man’s name has the typical structure
of Roman names, with the patronymic in the genitive — Cn(aei filius) — between
the praenomen Quintus, the nomen Arrius and the cognomen Mansuetus, all fol-
lowed by the origo Cluniensis. Probably this Quintus Arrius Mansuetus is the same
as the one who appears dedicating another slab to his mother Attilia Ammio
daughter of Gaius from Clunia, who died in the same place at the age of seventy
(no. 40). The dead woman’s native name is set out in a format close to the
32
De Palol and Vilella, Clunia II, no. 36. Intercatia, a town of Vaccaei was within the
conventus of Clunia; cf. TIR , K-30, Madrid, s.v. ‘Intercatia’.
33
De Palol and Vilella, Clunia II, no. 218: a funerary inscription, now lost. Vindeleia, a town
of Autrigones was within the conventus of Clunia; cf. TIR , K-30, Madrid, s.v. ‘Vindeleia’.
34
De Palol and Vilella, Clunia II, no. 35.
35
De Palol and Vilella, Clunia II, nos 61 and 102.
36
De Palol and Vilella, Clunia II, no. 12. The town of Brigaecini belongs to Conventus
Asturum; cf. TIR , K-30, Madrid, s.v. ‘Brigaecium’.
37
De Palol and Vilella, Clunia II, nos 13–14.
38
G. Baños Rodríguez, Corpus de Inscricións romanas de Galicia, vol. II: Provincia de
Pontevedra (Santiago, 1994), pp.101–65.
Page 214
Ordernr. 061297 P1
standard Roman name structure: a twofold personal name plus an abbreviated

affiliation, C(ai) f(ilia). Perhaps this Q(uintus) Arrius Mi(—) belonged to the
same family. He dedicated a funerary inscription to his wife Valeria Alla daughter
of Titus from Clunia who died in the same place at the age of twenty (no. 48). In
this case, the dead woman’s name is half way between the indigenous and the
Roman structure, being composed of two elements, an affiliation and the origo.
The people who commissioned those slabs all expressed their provenance from
Clunia. By comparison with other slabs from the same place, it has been suggested
that this group bears witness to a community of Cluniensis living in Vigo.39
Although the raw material was local granite, the slabs show a remarkable similarity
to others from Clunia and Lara:40 large stone blocks rectangular in shape and
rounded at the top; the epigraphic field is located lower down the stone, which is
crowned by a semi-circular space, outlined with decorative fluting. According to
the editors, ‘the typology of the Vigo slabs is clearly related to Galician-Roman
slabs; they were made by Galician-Roman craftsmen for a group of people with
different and more refined tastes. The Clunians brought their own epigraphic
design; they adopted the Galician-Roman type of monument while impressing
their own style upon it.’41
In Vicus Spacorum, we have a group of immigrants from Clunia in the far end
of Callaecia. They were not the only ones. Some inscriptions from the town of
Capera, far away in the province of Lusitania,42 mention Clunienses;43 one is par-
ticularly detailed: C. Caelius Pater/nu(s) Gal(eria) Cluniensis / an(norum) XXV
H(ic) S(itus) E(st) S(it) T(ibi) / T(erra) L(evis) Vicinia Cluniensium F(aciendum)
C(urauit).44 The vicinia Cluniensium placed and paid for an epitaph for a Roman
39
One of the slabs (no. 42) bears no mention of the dead person’s origin, but the formula*
peregr{a}e indicates a foreign resident, an alien (peregrinus) who died away from home.
40
J. A. Abásolo Álvarez, Epigrafía romana de la región de Lara de los Infantes (Burgos, 1974).
41
Baños Rodríguez, Corpus de Inscricións romanas de Galicia, pp.352–53. This demonstrates
that epigraphic usages obey deeper historical conventions than the simple fashions offered by
travelling workshops.
42
J. Gómez-Pantoja, ‘Historia de dos ciudades: Capera y Clunia’, in Économie et territoire en
Lusitanie romaine, ed. J.-G. Gorges and F. Rodríguez Martín, Collection de la Casa de Velázquez,
65 (Madrid, 1999), pp.91–108.
43
CIL, II, 818–20 and 822.
44
‘C. Caelius Paternus, of the voting-tribe Galeria, Clunian of 25 years. He lies here, may the
earth lie lightly upon you. The vicinia of Clunians took care of erecting it (the epitaph)’: CIL, II,
821. Hübner interpreted vicinia as a personal name.
Page 215
Ordernr. 061297 P1
200 Adela Cepas
citizen from Clunia, who had died at Capera. The meaning of the word vicinia in
this context is uncertain; it might be interpreted as a town quarter, a vicus, village
or hamlet, but, whatever the meaning, it points to a community of immigrants
from Clunia.
A Glimpse of Local Communities Below the City Level
We have seen so far that the discourse of civitas membership, as expressed in both
official and private inscriptions, dominated what little evidence we have of how
political identities worked at a local scale. The city was explicitly represented in
recorded official acts. Conversely it never featured as an attribute of the local
population while in Clunia, but was stated for those acting away from their city.
The implication is that, except for foreigners or passers-by, within the territory of
Clunia, the only accepted political denomination was the city itself, which need
not be expressed. This leaves us with the problem of the native groupings. If they
were taken to mean lesser local communities we could perhaps have a second layer
of community membership, but this is far from certain. It seems rather that those
terms referred to different, non-territorial kinds of social ties. Mention of those
native terms in the inscriptions does not seem to have competed with the city’s in
terms of political identities. City membership remained the only one that
mattered.
This is what the powerful discourse of social unanimity emanating from the
Roman municipal system proclaimed. Now, is that everything that existed by way
of political identities? The last inscription I would like to discuss seems to tell a
different story. This is a fragment of a bronze tablet measuring 13.8 x 14.7 cm
found in the village of Peralejos de los Escuderos (Soria). The piece originally had
a moulding fixed to the frame, of which only the left-hand side remains. The holes
for the nails, which must have fixed it to the wall, possibly in the forum of Termes,
are preserved in the frame. Its somewhat irregular script has been dated to the
second century AD. The inscription45 runs as follows: — / [—] ADIT[u et /
45
First published by A. D’Ors, ‘Un nuevo dato para la historia de la llamada Termancia’, in
Estudios dedicados a Menéndez Pidal, 3 vols (Madrid, 1951), II, 567–81, and Epigrafía Jurídica,
p.375, no. 25; Année Épigraphique (1953), no. 267, whose reading I follow; A. Jimeno, Epigrafía
romana de Soria (Soria, 1980), pp.160–61, no. 133; L. A. Curchin, ‘Vici and Pagi in Roman
Spain’, Revue d’Etudes Anciennes, 87 (1985), 327–43 (p. 331, no. 11); F. Beltrán Lloris, ‘Una
variante provincial del hospitium: pactos de hospitalidad y concesión de la ciudadanía local en la
Hispania Tarraconense’, in Epigrafía y Sociedad en Hispania durante el Alto Imperio: estructuras
Page 216
Ordernr. 061297 P1
Photo 9.1. The inscription from Peralejo de los Escuderos (Soria).

Photo: Museo Arqueológico Nacional, Madrid, Spain.
omnibus s]uis · ornament[is] populo · Termestino d(e) · s(ua) · p(ecunia) /5 f(acien-
dum) · c(urauerunt) · Dercinoassedensibus / vicanis · Cluniensium lib/eris · posteris-
que · eorum se/natus · populusque · Termestin/us · concessit · ut · eodem · iure es/10sent
· Termes · quo · ciues · Term/estini · IIIIuiris · L(ucio) · Licinio · Pilo / M(arco) ·
Terentio · Celso · L(ucio) · Pompeio / V(itulo) · T(ito) · Pompeio / Raro.46
There are four parts to the text. The first, which has been lost, would have
given the consular date as occurs in some documents of this kind. The second,
y relaciones sociales (Madrid-Alcalá de Henares, 2000), ed. S. Armani, B. Hurlet-Martineau and

A. U. Stylow, Acta Complutensia, 4 (Alcalá, 2003) pp.35–62 (p.44).
46
‘(The Dercinoassedenses) took care of having built and paid for at their own expenses (some
public work with an entry? and) all its ornaments for the People of Termes. The Senate and the
People of Termes granted the Dercinoassesenses, who lived in a vicus that belonged to the city of
Clunia, to their children and coming generations, the same law, whenever they were in the territory
of the city of Termes, of the citizens of Termes. (Signed by) L. Licinius Pilus, M. Terentius Celsus,
L. Pompeius Vitulus (and) T. Pompeius Rarus, members of the body of the four magistrates.’
Page 217
Ordernr. 061297 P1
202 Adela Cepas
partially missing, describes some type of work, perhaps public, that the Dercinoas-
sedenses had financed for the city of Termes.47 In the third, it is stated that the
Senate and the People of the city of Termes have conceded to the Dercinoasseden-
sibus themselves and their descendents to have the same law of the citizens of
Termes whenever they were in the territory of the civitas of Termes. Finally, the
document is validated with the names of the four magistrates, surely from Termes.
Usually this text has been interpreted as if the local citizenship had been
granted to be enjoyed anywhere at anytime. There are three references to Termes
and to the cives termestini. The first (line 8) refers to the Senate and the People;
the second (lines 8–9) to the law of the city; and between them (beginning of line
8) is a third mention: Termes. I think that the latter could be interpreted as a
locative (not declinable) referring to the time and space in which local citizenship
could be used: only while in the territory of the civitas of Termes. Therefore, the
locative Termes must have been used to express the terms in which the agreement
could be enforced;48 otherwise this third reference would have been redundant.49
Termes, an Arevacan city like Clunia, lies some 50 km to its south (Fig. 9.2).
The inscription was found in a modern village 5 km south-east from Termes,
which suggests that it very probably was carried there from the city ruins. We do
not know whether the Dercinoassedenses paid for the building of some public work
in Termes and as a result were welcomed into the city on equal legal terms as their
citizens or if things happened the other way around, but the order of the docu-
ment suggests that the initiative came from the first party. In whichever case, the
text records a treaty or agreement between two communities, cives and vicani,
which was settled on juridical terms, signed by the magistrates and put on the wall
of a public building, as was the custom in Roman towns. Although we only have
one copy of the document, possibly the one kept in Termes, the other party very
probably also had its own.
Unfortunately, the vicus Dercinoassedensis only occurs in this text, and its loca-
tion is unknown; however there is some evidence to help pin it down on a map of
pre-Roman Celtiberia. The term Dercinos is Celtiberic. While the genitive Asedi
appears on an inscription from the Lisbon area, its first part recalls a personal
47
Public works that D’Ors, Epigrafía Jurídica, p.376, suggests as being public baths.
48
D’Ors, Epigrafía Jurídica, p. 375, corrected Termes by the genitive (?) Termis. Jimeno,
Epigrafía romana de Soria, p.160, translates it as ‘belonging to Termes (with the same law as the
citizens of Termes)’; and Beltrán Lloris thinks that the term seems to be not declinable, ‘Una
variante provincial del hospitium’, p.44, n. 28.
49
Scribal mistakes are frequently attested on inscriptions, but less so in texts of this nature.
Page 218
Ordernr. 061297 P1
Figure 9.2. Detail of the Clunia area (after TIR, K-30, Madrid).
Page 219
Ordernr. 061297 P1
204 Adela Cepas
name, Dercinos — terkinos in the Iberic writing — five times mentioned in one of
the bronze plaques of Botorrita (Teruel);50 it also bears some similarity with the
place-name Dercenna, located by Martial51 in the Celtiberic region of Bilbilis and
with a native divinity called Dercetius, recorded in a second-century votive
inscription from near modern San Millán de la Cogolla.52 Dircetius seems to be a
mountain name that later passed on to the monastery of San Millán de la Cogolla,
occasionally named Dercensis until the tenth century.53
Now, wherever it lay, what is most remarkable in this text is that one of the
parties was a vicus of Clunia, thus a lesser settlement subordinated to it, but it was
able to act legally in a perfectly autonomous fashion in no less a matter than nego-
tiating a second citizenship with another community of a higher rank, a civitas.
Moreover, it had control of its own finances, thus coming close to a res publica. All
this suggests an almost urban character, even if under the cover of the civitas of
Clunia.54
The inscription says nothing about why these vicani would pay to be treated
on legally equal terms with the cives termestini, or what it meant to be subjected
to the same rights and obligations (munera).55 On the other hand, the relationship
established between them is not of the type usually shown on a tessera* hospitalis
or a tabula patronatus not only because the usual words, hospitium or patronatus,
do not appear (at least not in the preserved text) but also because of the content
50
F. Beltrán Lloris and others, El tercer bronce de Botorrita (Contrebia Belaisca) (Zaragoza,
1996); Beltrán Lloris, ‘Una variante provincial del hospitium’, p.44, n. 29. The Celtiberi took the
semi-syllabic writing from the eastern Iberi as a means of expressing their own Celtic language; as
the Roman power gained strength they moved to Latin.
51
Martial, Epigramas, ed. M. Dolç (Barcelona, 1949–50), I. 49.
52
Found at San Cristóbal on the slopes of Monte Castillo. Mª. L. Albertos, ‘El Culto a los
montes entre los Galaicos, Astures y Berones y algunas de las deidades más significativas’, Estudios
de Arqueología Alavesa, 6 (1974), 47–157; U. Espinosa Ruiz, Epigrafía romana de la Rioja
(Logroño, 1986), pp.59–60, no. 40; TIR , K-30, Madrid, s.v. ‘San Cristóbal (San Millán de la
Cogolla)’.
53
(Madrid, 1943), p.16. The area of San Millán de la Cogolla — where the Roman inscription and
the seventh-century place-name are documented — is near to the Sierra de la Demanda, the likely
boundary between the conventus of Clunia and Caesaraugusta.
54
Already pointed out by Beltrán Lloris, ‘Una variante provincial del hospitium’, p.44, and
by Curchin, ‘Vici and Pagi’, p.335.
55
G. Pereira Menaut, ‘Che cos’è un munus?’, Athenaeum, 1 (2004), 169–215.
Page 220
Ordernr. 061297 P1
of the text itself.56 Either they were often present in the city of Termes — whether
for a long stay or just passing through its territory57 — or there was a stable com-
munity of immigrants there, as was the case of Clunians in Vicus Spacorum or
Capera.58
Roman inscriptions from Spain sparingly reveal hints of vici, pagi and castella
somehow acting in an autonomous fashion.59 Surely the reality of local politics
operated in ways rather different from those that the discourse emanated from
major power centres would have us envisage. The very presence of such communi-
ties of sub-urban rank is a fine reminder that the Roman cities were no more a
resource for the Roman administration to control its realm than a device by which
urban — mainly native — elites reinforced their control over an enlarged terri-
tory. This surely could not happen without conflict at the local level, but we
hardly ever see traces of it.
Nevertheless, such a system of local domination could not stay unaltered across
the changes that affected the region’s Roman towns from the third century. The
dramatic changes evidenced in Clunia’s archaeological record probably point to
parallel developments in its hinterland. In this respect, the archaeological visibility
of the different kinds of sites may well have misled scholars to compare the decline
in urban structures to the splendour of rural villa-sites. This may be a false equa-
56
Generally the Latin tesserae and pacts of hospitality and patronage found in Spain sanction
individual agreements between two parties: Romans and provincials; or a city and a person who
is granted the local citizenship; or two Roman towns. It is worth noting that these sorts of
documents are only found in Italy, Spain and North Africa; F. Beltrán Lloris, ‘La hospitalidad
celtibérica: una aproximación desde la epigrafía latina’, Palaeohispanica, 1 (2001), 35–62, and
F. Beltrán Lloris, ‘Los pactos de hospitalidad de la Hispania Citerior: una valoración histórica’, in
La Península Ibérica hace 2000 años, ed. L. Hernández Guerra, L. Sagredo San Eustaquio and
J. Mª. Solana Sáinz (Valladolid, 2001), pp.393–99.
57
In search of pasture lands, as suggested by J. Gómez-Pantoja, ‘Pastio agrestis: Pastoralismo
en Hispania romana’, in Los rebaños de Gerión: Pastores y trashumancia en Iberia antigua y
medieval, ed. J. Gómez-Pantoja, Collection de la Casa de Velázquez, 73 (Madrid, 2001),
pp.177–213. E. Haley, Migration and Economy in Roman Spain (Barcelona, 1991), p.106, thinks
that economic reasons of uncertain nature led them to look for links with the people of Termes.
58
There are other places where Clunienses — as well as Uxamenses — appear. Of the seven
hundred inscriptions found in Iberia mentioning foreigners, those pertaining to these two cities
constitute an eighth of the total and there are more than twice as many as those from any other
place, Haley, Migration and Economy, pp.87–88.
59
Curchin, ‘Vici and Pagi’; P. Le Roux, ‘Vicus et castellum en Lusitanie sous l’Empire’, Studia
Historica: Historia Antigua, 10–11 (1992–93), 151–60.
Page 221
Ordernr. 061297 P1
206 Adela Cepas
tion. The case of Clunia may be better explained by pointing out that, from the
start, the city’s prominence was largely based on the position it was granted within
the Roman administrative network. Without the latter, it seems to have lapsed
back among its peer settlements, even losing out to Vxama in the competition to
be an episcopal see in the Visigothic period.
More importantly, the city’s decline may well have meant a decrease in its
ability to hold control of a territory where, as we have seen, other forms of
political allegiance existed, but rarely emerged from below the cover of the
centrally enforced discourse of local citizenship. The case of the vicus Dercinoasse-
densis (even if we cannot locate it) may parallel other developments in the
archaeological record. All across the east of the Duero plateau there is a repeated
pattern of hilltop sites* which were occupied in the late Iron Age but have no
early Roman phase. Many of them, though, were reoccupied in the late empire,
often in the late fourth century, that is, before the collapse of Roman authority
and the beginning of the fifth-century invasions. There are a number of such sites
in Clunia’s vicinity, and probably more are to be discovered in the future. Sites as
La Yecla, Peña de Nuestra Señora (surrounds of La Yecla), El Castillo, El
Castillejo, Cabeza de San Vicente (surrounds of Castrovido), etc. fit within this
pattern.60
The reoccupation of old native hilltop sites is yet ill understood, but it should
be ascribed neither to a ‘Celtic revival’ nor to random selection of prominent loca-
tions that just happened to have a pre-Roman phase. The evidence discussed in
this essay suggests that, below the layer of urban territories and government, lesser
communities existed throughout the Roman period61 whose visibility was largely
blocked off by the centre’s prominent material culture and political discourse.
Their material structures, as well as their relationship to both old hill-fort sites
and rural villa-sites, still awaits investigation. The possibility that their archaeo-
60
J. A. Abásolo, and R . García Rozas, Carta arqueológica de la provincia de Burgos: Partido
judicial de Salas de los Infantes (Burgos, 1980), pp.29–30; J. Escalona Monge, Sociedad y territorio
en la Alta Edad Media Castellana: La formación del alfoz de Lara, British Archaeological Reports,
International Series, 1079 (Oxford, 2002), pp.53–54.
61
All the inscriptions discussed in this essay are dated to the first and early second centuries
on the basis of palaeographic and stylistic characteristics and the internal structure of the texts.
However, the ‘epigraphic habit’ of Romans and provincial as defined by MacMullen (‘The
Epigraphic Habit in the Roman Empire’, American Journal of Philology, 103 (1982), 233–46)
seems to come to an end by the mid-third century (S. Mrozek, ‘À propos de la répartition
chronologique des inscriptions latines dans le Haut-Empire’, Epigraphica, 35 (1973), 114).
Page 222
Ordernr. 061297 P1
logical emergence was just another factor in the process of urban decline and
territorial fragmentation should also be considered.
If confirmed, this would not only provide an interesting insight into the long-
term evolution of settlement patterns and local communities between the Roman
and early medieval period, but also a deeper historical background for the pro-
cesses of early community formation investigated in other essays in this volume.62
62
See especially Martín Viso and Halsall, this volume. I am very grateful to José Joaquín
Caerols (Universidad Complutense, Madrid), Armin Stylow and Helena Gimeno (Centro CIL
II, Universidad de Alcalá, Madrid) for their help with the interpretation of some inscriptions.
Also to Julio Escalona (CSIC, Madrid) for reading and commenting on an earlier draft. This essay
was written during a fellowship granted by the Comunidad Autónoma de Madrid.
Page 223
Ordernr. 061297 P1
Page 224
V ILLAS, T ERRITORIES AND C OMMUNITIES

IN M EROVINGIAN N ORTHERN G AUL
Guy Halsall
T
his essay confronts the definition of the Merovingian villa: how the villa
functioned within local politics, and how it worked in relationships
between insiders and outsiders. Revisiting the archaeological and docu-
mentary data from the Merovingian region of Metz discussed in a previous
monograph (Fig. 10.1),1 it attempts to refine earlier hypotheses in the light of
more recent archaeological work in this area and elsewhere in Gaul and the post-
imperial West.
The Villa
The Roman Villa
In documentary sources, a villa* is generally a country house, a country seat or a

farm.2 The term fundus is more usual for an estate, and it has been suggested that
the use of villa to mean fundus was a late antique development.3 Archaeologically,
1
G. Halsall, Settlement and Social Organization: The Merovingian Region of Metz (Cam-
bridge, 1995). The research for this monograph was completed in mid-1993 and a number of
important works published in 1992–93 were not yet available to me.
2
Where the word ‘villa’ is placed in italics I refer to its technical usage in contemporary
documents. Elsewhere, I have, for clarity, coined terms such as ‘villa-community’ and ‘villa-region’
to distinguish them from each other and from the actual buildings of the villa proper.
3
G. Ripoll and J. Arce, ‘The Transformation and End of Roman villae in the West
(Fourth–Seventh Centuries): Problems and Perspectives’, in Towns and their Territories between
Page 225
Ordernr. 061297 P1
210 Guy Halsall
I shall understand a villa to be a residence that was a visible focus for the expen-
diture of surplus extracted by its owner, constructed to Roman or Romanizing
stylistic norms. We need not assume that the villa was necessarily the dwelling of
the owner of the directly surrounding lands — that an estate belonging to the villa
immediately adjoined it. Surplus could have been extracted elsewhere or by other
means. Nor need we assume that, where it was, the landowner necessarily dwelt
in the villa most or all of the time. That said, I shall assume that most villas under
analysis in northern Gaul provided, in some sense or other, a focal point for the
local community. Given the archaeology of the north Gallic villa, I shall also
assume that in this region most were centres of working, farming establishments
with a tenurial relationship of some sort with the lands round about.
North of the Seine, villas were generally of small to middling size. There were
also many fewer after the third century, though the extent of numerical decline
remains debatable. The degree of complete abandonment in the late third century
has probably been overestimated.4 Nevertheless, depending upon the region under
discussion, between a quarter and half of the villas in existence at the start of the
third century had apparently been deserted by the beginning of the fourth.5 The
process of abandonment continued, albeit unevenly, until the troubles of the
decades around 400 killed off the survivors. More optimistic survival rates are
often based simply upon continued occupation of a site whereas a key tenet of this
study will be that the nature of occupation is crucial. Fifth-century and later
occupation sometimes takes the form of small wooden buildings, indistinguish-
able in material cultural terms from other sites in the settlement pattern and
employing traditional indigenous architectural forms.6 This surely suggests that
Late Antiquity and the Early Middle Ages, ed. G. P. Brogiolo, N. Christie and N. Gauthier
(Leiden, 2000), pp.63–114 (pp.63–66). Their discussion of the documentary usage of villa is
based largely upon secondary sources and is somewhat problematic.
4
For example, Halsall, Settlement, pp.176–78, based on problematic figures given by E. M.
Wightman, Gallia Belgica (London, 1985), p.259.
5
P. van Ossel Établissements ruraux de l’antiquité tardive dans le nord de la Gaule, 51 e
Supplément à Gallia (Paris, 1992), pp. 61–84. The discussion of villa survival in C. R . Whittaker
and P. Garnsey, ‘Rural Life in the Late Roman Empire’, in The Cambridge Ancient History, vol.
XIII: The Late Empire, A .D . 337–425, ed. A. Cameron and P. Garnsey (Cambridge, 1998),
pp.277–311 (p.297), seems unduly optimistic.
6
For example, Neerharen-Rekem (Belgium): van Ossel, Établissements ruraux, pp.297–300.
At Hoogeloon the stone building was replaced around 200 by a Wohnstallhaus of indigenous
tradition: M. Carroll, Romans, Celts and Germans: The German Provinces of Rome (Stroud, 2001),
p.65.
Page 226
Ordernr. 061297 P1
VILLAS, TERRITORIES AND COMMUNITIES 211
Figure 10.1. The Region of Metz.
Page 227
Ordernr. 061297 P1
212 Guy Halsall
in the post-imperial period there existed a different social and economic situation,
different possibilities for and extent of surplus extraction, and a different set of
relationships between the owners of these sites and the inhabitants of the other
settlements in the area from that made manifest earlier by large stone-built, tile-
roofed buildings with hypocausts.
There were of course exceptions. Around Trier, the aristocracy was wealthier
and more powerful as well as having a direct linkage to the emperor and his court
and resources during the fourth century. Unsurprisingly, therefore, villas there
can be very substantial.7 Oddly, it is also in this region that evidence of late third-
century destruction is best known.8 The patchy data from the region of Metz
present some similarities to those from around Trier. Survival of villas beyond the
third century was comparatively good, but those that did continue were generally
small, the most important exception being the great villa at St-Ulrich.9 It is diffi-
cult to say how big the estates centred on these villas were. Marcel Lutz believed
that the territory centred on the exceptional villa at St-Ulrich was extensive, but
the scale of villa-buildings does not generally suggest huge estates. In modern
communes, there is usually only one place-name with an ending believed by place-
name specialists to derive from Latin estate names (most obviously -y from -iacum,
a common ending for Roman estate names10). Communes here usually encompass
a half-dozen or so square kilometres, perhaps indicating the size of a region of
which a villa was focus. It should not be assumed that a villa proprietor owned all
of this land, and it might also be that these place-names are post-Roman.
This archaeological data does not look like the material cultural signature of
a society based around a massively wealthy aristocracy dominating an impover-
ished dependent peasantry or (even more disputably) rural slavery, such as is still
often assumed to have characterized late Roman social relations. We might admit
that there could be, and almost certainly were, absentee landlords resident else-
where — in the south or even further afield — who owned properties in northern
Gaul. Some villas might represent hunting lodges or similar. Nevertheless the
third- and fourth-century changes underline that the conventional picture of late
Roman social structure cannot simply be assumed to have applied in northern
Gaul.
7
For example, Konz (Kr. Trier-Saarburg): van Ossel, Établissements ruraux, pp.244–49.
8
Van Ossel, Établissements ruraux, p.70.
9
M. Lutz, ‘Le domaine gallo-romaine de Saint-Ulrich (II)’, Gallia, 30 (1972), 41–82.
10
Sidonius Apollinaris’s estate was called Avitacum for example.
Page 228
Ordernr. 061297 P1
The imperial link is crucial. Even with the creation of the field armies, spread
more evenly across, and deeper into, the provinces,11 northern Gaul (including the
Germaniae) was still the location of large concentrations of troops, all requiring
food and provisions. There is considerable archaeological evidence of the militari-
zation of the region, such as large, clearly fortified rural granaries.12 Their scale,
compared with that of ‘private’ building, has led to the reasonable assumption
that they are connected with the supply of the armies and the bureaucracy.
The archaeological evidence from the critical period between 383 and 421 also
suggests that the local aristocracy was very militarized. Burials that appear in the
region during those decades reveal local elites demonstrating their power through
the use of weaponry and badges of imperial, probably military, office — official
belt-sets and brooches.13 The reduction in the numbers of northern Gaulish villas
might manifest an increasing militarization of society. It might more plausibly
reflect an imperial takeover of lands and their produce to feed the Rhine garrisons,
the praesental field army and the central imperial bureaucracy at Trier. This pro-
cess might be revealed by the Langmauer, the long wall surrounding an estate near
Trier known, from epigraphic evidence, to have been constructed by troops from
a legion of Primani.14 Such estates could have been granted, through emphyteutic
leases (long-term leases, carrying obligations to improve the land), to imperial
officers. Their yields might alternatively have been granted on a short-term basis
to imperial servants, the comparatively rapid turnover of recipients of such
11
D. Hoffmann, Das Spätrömische Bewegungsheer und die Notitia Dignitatum, 2 vols,
Epigraphische Studien, 7.1 and 7.2 (Düsseldorf, 1969–70); R . S. O. Tomlin, ‘The Army of the
Late Empire’, in The Roman World, ed. J. Wacher (London, 1987), pp.107–23; H. Elton,
Warfare in Roman Europe, 350–425 (Oxford, 1996); A. D. Lee, ‘The Army’, in Cambridge
Ancient History, vol. XIII, ed. Cameron and Garnsey, pp.211–37.
12
P. van Ossel, ‘Insécurité et militarisation en Gaule du Nord au Bas-Empire: L’exemple des
campagnes’, Revue du Nord, 77 (1995), 27–36; H. Bender, ‘Archaeological Perspectives on Rural
Settlement in Late Antiquity in the Rhine and Danube Area’, in Urban Centers and Rural Con-
texts in Late Antiquity, ed. T. S. Burns and J. W. Eadie (East Lansing, MI, 2001), pp.147–61.
13
G. Halsall, ‘The Origins of the Reihengräberzivilisation: Forty Years On’, in Fifth-Century
Gaul: A Crisis of Identity?, ed. J. F. Drinkwater and H. Elton (Cambridge, 1992), pp.196–207;
G. Halsall, ‘Archaeology and the Late Roman Frontier in Northern Gaul: The So-Called
Föderatengräber Reconsidered’, in Grenze und Differenz im früheren Mittelalter, ed. W. Pohl and
H. Reimitz (Vienna, 2000), pp.167–80. These burials should not be associated with Germanic
settlers.
14
H. Cüppers, ed., Trier: Kaiserrezidenz und Bishofssitz: Die Stadt in spätantiker und
frühchristlicher Zeit (Mainz, 1984), pp.288–91.
Page 229
Ordernr. 061297 P1
214 Guy Halsall
patronage explaining why this surplus was no longer turned into permanent high-
status buildings (villas) on the estates.
Whatever the case, we should stress the apparent stability of local society for
much of the fourth century; the importance of the empire and its structures in
ensuring such stability, either through maintaining social order amongst a broad
and not very wealthy local ‘squirearchy’ by law, office and patronage, or in main-
taining absentee landlords’ security of tenure of far-flung estates, or in providing
frameworks for the extraction of surplus, markets and so on; and the importance
of the fiscal role of land and produce. This last point and the fact that the local
community was keyed into, and perhaps entirely dependent upon, a wider set of
political structures is especially important.
The Merovingian Villa
Roman villa-complexes (as defined above) are extinct by c.450 in this part of the
world, and usually earlier. That leaves the thorny issue of where exactly early
Merovingian high-status settlements were, an issue to which I have at best only
tentative answers.15 Nonetheless a few people living in huts amidst ruined stone
buildings, as at Nennig in the southern Triererland,16 does not, in any meaningful
sense, a villa make. There is, however, general continuity of the settlement pattern
between the late Roman and Merovingian periods; the distribution map of Mero-
vingian cemeteries is much the same as that of late Roman settlements.17 There
is some contraction but this seems to have taken place earlier during the late Ro-
man period. Stray and chance finds suggest a close relationship between substan-
tial Roman buildings — probably (though not certainly) villas — Merovingian
cemeteries, churches and modern villages.18 The bulk of known Merovingian
15
People wore their wealth, as discussed below; perhaps powerful people lived in former
Roman higher-order settlements: vici, castra, castella* or towns.
16
K. Böhner, Die fränkischen Altertümer des Triererlandes, 2 vols, Germanische Dennkmäler
der Völkerwanderungszeit, Serie B, Band 1 (Berlin, 1958), II, 88–91. Van Ossel, Établissements
ruraux, pp.383–84. Above, pp.210, 212.
17
F. Stein, ‘Die Bevölkerung des Saar-Mosel-Raumes am Übergang von der Antike zum
Mittelalter: Überlegungen zum Kontinuitätsproblem aus archäologischer Sicht’, Archaeologia
Mosellana, 1 (1989), 89–195, Abb. 5 and 6.
18
Modern survey in the département of Aisne suggests similar patterns: C. Haselgrove and
C. Scull, ‘Early Medieval Rural Settlement in North-East France: Field Survey Evidence from the
Aisne Valley, Picardy’, in Early Medieval Europe 1992: A Conference on Medieval Archaeology in
Page 230
Ordernr. 061297 P1
cemeteries are close to modern settlements,19 suggesting a general continuity from

Roman to modern of occupation in the area of some modern villages. This con-
clusion, proposed from my systematic survey of the extant data in 1986–93, has
recently been confirmed by a detailed study of the commune of Basse-Yutz.20
Here clear evidence was revealed of continued occupation in the general area of
those Roman settlements (at Haute- and Basse-Yutz) that continued to be occu-
pied into the late Roman period, whereas early Roman sites that did not survive
into the fourth century showed no trace of post-Roman settlement. This means
neither that the Merovingian settlement pattern entirely underlies the modern,
nor that Merovingian settlements excavated away from modern villages are ‘failed
settlements’.21 What it does suggest is that one element, and I suggest a prestigious
element, of the settlement pattern underlies the modern network. The Yutz study
supports this postulation in that a royal villa existed there by the mid-ninth
century. The fractious sons of Louis the Pious met there in 844 for one of the con-
ferences aimed, with general lack of success, at ending their squabbling.22
With no archaeologically attested Merovingian villas, we must ask what Mero-
vingian writers mean when they refer to villae. There are very few written sources
Europe, 21 st–24 th Sept 1992 at the University of York: Pre-printed Papers, 8 vols (York, 1992), VIII,
207–12. Cf. E. Louis, ‘A De-Romanised Landscape in Northern Gaul: The Scarpe Valley from
the Fourth to the Ninth Century AD ’, in Recent Research on the Late Antique Countryside, ed.
W. Bowden, L. Lavan and C. Marchado (Leiden, 2004), pp.479–504.
19
M. Durand, Archéologie du cimetière médiévale au sud-ouest de l’Oise: relations avec l’habitat
et évolution des rites et des pratiques funéraires du VIèm e au XVIèm e siècle, Revue Archéologique de
Picardie, numéro spéciale (Amiens, 1988); M. Durand, ‘Cimetières et habitats: l’émergence du
village au sud-ouest du département de l’Oise’, Revue Archéologique de Picardie, 3–4 (1988),
153–60. E. Zadora-Rio, ‘The Making of Churchyards and Parish Territories in the Early
Medieval Landscape of France and England in the 7th –12 th Centuries: A Reconsideration’,
Medieval Archaeology, 47 (2003), 1–19 (esp. pp.8–9).
20
J.-M. Blaising, ‘Yutz (57), archéologie d’un terroir des âges de metaux aux XIX e siècle’,
Archaeologia Mosellana, 4 (2002), 185–217 (esp. p.213). Halsall, Settlement, p.200, discussing
Yutz, repeats the then general assumption that Basse-Yutz was a larger agglomeration, or vicus (the
evidence discussed is that, specifically, from Basse-Yutz though this is not made clear). Blaising,
‘Yutz’, stresses the lack of evidence for this assumption.
21
H. Hamerow, Early Medieval Settlements: The Archaeology of Rural Communities in North-
West Europe 400–900 (Oxford, 2002), p.106, recently and rightly takes to task my slack use of
that phrase: Halsall, Settlement, p.184.
22
Most chronicle sources state that this meeting took place across the Moselle in Thionville
but the acts of the ecclesiastical council place it in Iudicium. Die Konzilien der karolingischen
Teilreiche, 843–859, ed. W. Hartmann, MGH Concilia, 3 (Hannover, 1989), p.29.
Page 231
Ordernr. 061297 P1
216 Guy Halsall
from this region before the third quarter of the seventh century so any usage
uncovered might represent the end of a process. Traditionally, the villa, in docu-
mentary sources, is an estate. The word does have this sense in some documents,
especially when referring to royal estates or gifts, but it is not its only meaning.
The cartulary* of Wissembourg Abbey, founded in the mid-seventh century,23 is
well known as one of the earliest repositories of non-royal documents, giving an
insight into the late Merovingian landscape. In its documents a villa is a unit of
land, within which several people could have estates. Until the 780s villa was
seemingly a synonym for fines* (bounds; limits) or its Germanic counterpart
marca* (whence the English ‘marches’). Rarely is more than one of these words
found in the same document, suggesting that one could stand for the other. With
the exception of the villa of Waldhambach, which the charters repeatedly define
as between the rivers Mittelbrunn and Eichel (still respectively the southern and
northern bounds of the commune of Waldhambach), there are no boundary
clauses for these villae. Nor are they measured; people seem to have known where
their edges were. Mostly these are quite extensive grants of land.
From the 780s the documents change to referring to grants in villa vel in fines
or vel in marca (in the villa or in the fines [or marca]).24 Such a shift is perhaps
adumbrated by an early eighth-century charter referring to the gift of a curtilis (small
farm) in Audoneuillare and fields and woods in the fine Auduninse.25 This suggests
that the villa (in this case incorporated in the place-name Audoneuillare) could
be distinct from the fines (and note that in this case the name Auduninse ends in
-ensis, the ending used in Latin names for regions — for example in the pagus Sali-
nensis, the administrative district, later county, around the River Seille). It might,
alternatively, be argued that the earlier usage in the Wissembourg cartulary im-
plies that villa and fines were clearly different, and hence no charter discusses the
two together, whereas the change after the 780s argues that the two had become
interchangeable. This is plausible but I prefer to interpret the change as showing
that hitherto interchangeable terms became distinguished from each other, not
least because some later documents specify villae within fines.26 Also from the
780s, the Wissembourg documents begin to refer to small, measured parcels of
23
Halsall, Settlement, pp.188–92. Traditiones Wizenburgenses: Die Urkunden des Klosters
Weissenburg, 661–864, ed. K. Glückner and A. Doll (Darmstadt, 1979).
24
Halsall, Settlement, p.190.
25
Wissembourg, charter 228, dated 705–06.
26
For example, Wissembourg, charter 217, dated 783; Wissembourg, charter 82, dated 786.
Page 232
Ordernr. 061297 P1
land and, whereas large grants bestow lands ‘in the villa or in the marca’, these
lesser grants are usually specifically located in either the villa or in the marca / fines.
Above all, the latter interpretation, of a distinction of villa from fines after 780,
is supported by developments in the charters of the monastery of Gorze. Gorze
was founded in the mid-eighth century and reformed in the tenth. Its charters’
information about landscape organization needs closer study and re-examination
in the light of John Nightingale’s study of the reformed monasteries in Lorraine.27
Nightingale’s detailed, scholarly and more specialist analysis of these charters
seriously questions the authenticity of the early Gorze documents, especially the
detailed and interesting charter no. 19.28 The latter is apparently a tenth-century
document wrongly dated to the late eighth century by the cartulary’s compilers
on the basis of a confusion of one King Charles (‘the Simple’) with another (‘the
Great’). Nevertheless, the indictional dates and other chronological indices care-
fully inserted by the later scribe are most consistent, with each other and with the
regnal years given in the charters. Leaving the clearly spurious charters aside, the
chronology of changes in the vocabulary used to describe settlements is also
remarkably consistent and seems unlikely to result from later forgery.
This chronology, moreover, ties in well with the Wissembourg documents.
Early, or at least putatively early, Gorze charters refer to lands in villae in much
the same way as the earlier Wissembourg charters. Then, from c.775, while a
change was taking place in the Wissembourg charters’ formulae*, they switch to
the formula ‘in villa N (usually a personal name in the genitive) vel in fines N
(ending in -iacum, the old Roman estate place-name termination, -ensis, the re-
gional ending, or, occasionally, both)’. Sometimes villa is replaced by its synonym
curtis (‘courtyard’ or farm, underlining that this element refers to a specific settle-
ment rather than a region). Villa and curtis have given rise to the large number of
27
Halsall, Settlement, pp.192–95. J. Nightingale, Monasteries and Patrons in the Gorze Reform
(Oxford, 2001), pp.25–29. Cartulaire de l’Abbaye de Gorze (Mettensia 2), ed. A. D’Herbomez
(Paris, 1898; introduction dated 1902). The possibilities for refined study are limited by the
destruction of the earliest (thirteenth-century) manuscript of the Gorze cartulary, upon which
D’Herbomez’s edition was based, during street-fighting in Metz in late 1944. The German
garrison made its last stand on the Île de la Préfecture, location of the city archives. Indeed, sent
out to negotiate the surrender was the archivist, none other than the celebrated Merovingian
historian, Eugen Ewig. The surviving manuscripts are both less complete and later (eighteenth-
and fifteenth-century respectively) but might profitably be examined more closely. D’Herbomez
apparently used neither.
28
Cited, alas, at length by Halsall, Settlement, p.193.
Page 233
Ordernr. 061297 P1
218 Guy Halsall
place-names ending in ‘-ville’ or ‘-court’ in the region.29 One might translate thus:
‘in so-and-so’s villa or within the bounds of that territory’. Three examples: ‘within
the bounds of “Ragnulfiacum” (an -acum name) or in that villa of “Ragnulf’s
Villa”’ (in Ragnulfiaga fine vel in ipsa villa Ragnulfivilla); ‘within the bounds of
“Lagbriacensis” (-acum plus -ensis) or in “Labrigia villa”’; or ‘within the bounds of
“Ingariensis” (an -ensis name) or in the villa called “Hingarigo Curtis”’.30 This lasts
up to the 850s before being replaced by in villa N vel in ipsa fines (‘in so-and-so’s
villa or within those bounds’). Again, though, while the Gorze cartulary contains
numerous boundary clauses and measured units of land, largely because the lands
donated tend to be small parcels, the boundaries of the fines are never defined.
Thus, in non-royal contexts, a Merovingian villa is essentially a small region.
‘Lagbriacensis’, for example, might equate with the commune of Labry. Bange
reached a similar conclusion about the meaning of villa in tenth-century Burgun-
dian documents.31 Early sixth-century Burgundian law already specified that a
villa could be understood in this way.32 One clause in the Frankish Pactus Legis
Salicae suggests the importance of the villa-community: the vicini (neighbours)
who can accept or reject newcomers into their community live in villae.33 Finally,
at one point in the so-called ‘feud’ of Sichar and Chramnesind (586–87), Chram-
nesind burnt Sichar’s house and those of all those who shared the villa (spoliatis
[. . .] domus omnes tam Sichari quam reliquorum qui participes huius villae erant).34
Within the villa-region the basic unit of settlement was the mansus*, sometimes
called by a latinized version of its Germanic equivalent, hoba* (Hof), in the Wis-
sembourg documents. Mansi could be quite large; in the Merovingian documents
from the Metz region they could be granted with lists of appurtenances.35 Only
29
See distribution map in Halsall, Settlement, p.11.
30
Cartulaire de l’Abbaye de Gorze, ed. D’Herbomez, charters 26, 28 and 37, dated 786, 788
and 795 respectively. None of these charters is specified as a forgery by Nightingale.
31
F. Bange, ‘L’ager et la villa: structures du paysage et du peuplement dans la région Mâcon-
naise à la fin du haut moyen âge (IX e– XIe siècles)’, Annales ESC, 41 (1984), 529–69.
32
Liber Constitutionum 38.4; Leges Burgundionum, ed. F. de Salis, MGH Legum Sect. I, 2.1,
(Hannover, 1892).
33
Pactus Legis Salicae 45, 102, ed. K.-A. Eckhardt, MGH Legum Sect. I, 4.1 (Hannover,
1962), pp.173–76.
34
Gregory of Tours, Histories, VII.47, ed. B. Krusch and W. Levison, MGH SRM, 1.1
(Hannover, 1951), p.367.
35
That mansi are found within villae poses problems for Ripoll and Arce’s statement (‘Trans-
formation and End of Roman villae’, p.66) that the mansus is the seventh-century equivalent of
the villa.
Page 234
Ordernr. 061297 P1
in royal documents can the villa be understood to mean a single property

(although this might contain several settlements). Merovingian kings habitually
grant our villa called N. Though few Merovingian royal charters survive for the
region under discussion in this essay it is true for those that do exist: villam
nostram nuncupante Marte (‘our villa called Mars [la Tour]’) for instance. This
usage does not seem specific to the region of Metz. In other documents, royal
villae are conceived of as single units of royal landed property.
From Roman to Merovingian Villa
How did the villa develop from a Roman farmhouse or estate into a Merovingian
region? With no Roman documents from the region the apparent transformation
in usage could simply reflect a change in the rates of documentary survival. Per-
haps a villa in Gaul was always a region; this would make a certain sense. A second
alternative would see the villa as a fiscal unit of assessment. Landowners were
responsible for their tax professio, and an estate and dependencies was a fiscal unit.
Jean Durliat has proposed this.36
A third possibility involves local politics. Early fifth-century villa abandon-
ment appears to have been brought about by the crisis of the empire. The end of
effective imperial presence in the north caused enormous stress, as one might
expect given the points made above. This was exacerbated by the fact that Roman
rule was not terminated neatly by rapid conquest or formal treaty. It was instead
a messy process of de facto abandonment as in Britain, where the results were
similar, and Spain, where they were not.37 Northern Gaulish social organization
had, as argued above, depended upon the empire’s presence. Without that, the
local ‘squirearchy’ was, like the British aristocracy but unlike the Spanish, unable
to fill the power vacuum. Without the support provided by the empire, local
power was up for grabs and expensively maintained from one generation to the
next. Such surplus as was acquired was dissipated in gifts to purchase or reinforce
support, leaving none for the maintenance of villa-buildings or industry.38 From
36
J. Durliat, Les finances publiques de Dioclétien aux Carolingiens (284–890), Beihefte der
Francia, 21 (Sigmaringen, 1990), pp.253–56. I do not accept all of Durliat’s argument.
37
For narrative of the end of empire in the west and a survey of its effects, see my forthcoming
Barbarian Migrations and the Roman West (Cambridge), chapters 6–9 and 11.
38
Decline might be overplayed — pottery was exchanged across large regions. Nevertheless
there was serious contraction. The range of decoration and forms of vessels are dramatically
Page 235
Ordernr. 061297 P1
220 Guy Halsall
the last quarter of the fourth century, evidence of occupation at Metz itself is very
slender; such evidence peters out at the vici*/castra of the region from c.400; and
the final phase of villa abandonment had been played out by c.425.39 Thus, socio-
political interaction retreated to a very local arena, the villa-region, which might
further explain the continuity of the unit and term villa, even as what lay behind
the word changed. The change in the nature of the villa might involve a mixture
of all the above factors.
Insiders: The Community
How did the Roman villa-buildings act as a focus for a community and how did
the community work to preserve the villa as a geographical/social unit? We might
explore this through the region’s cemeteries. Many Merovingian cemeteries are
adjacent to Roman villas.40 In the Metz region the latter are, as far as can be told
from the available data, generally those whose occupation continued after the
third century. Even if old tenurial relationships had broken down in the early fifth
century, the ruins of the villa-buildings seemingly continued to provide a focus for
important communal rituals. Perhaps the villa’s owners began this by staging im-
portant funerary rites at the site. The well-known lavish burial 319, for instance,
‘founds’ the cemetery at Lavoye, located by a villa occupied throughout the fourth
century.41 The important families of a community need not have lost their pre-
eminence but it was apparently much more open to threat and expensive to main-
tain. By c.525 whole communities were participating in the competitive ritual of
furnished burial. If funerary ritual was important in helping to define a commu-
nity, on the basis of its shared cemetery near the old Roman villa-buildings (which
might have given the community its toponymic identification — ‘the people of
such-and-such villa’), then this community seems to have been shaped by a central
reduced, and the fact that even these plain pots were sought over comparatively long distances
indicates some economic decline. Such basic wares seem hitherto to have been produced on villa-
estates*: M. Corbier, ‘Grande proprietà fondaria e piccole azienda: La Gallia settentrionale in
epoca Romana’, in Società Romana e Impero Tardoantico, vol. III: Le merci, gli insediamenti, ed.
A. Giardina (Rome, 1986), pp.687–702, with notes at pp.897–903 (p.699). Though we must
agree to differ, I am immensely grateful to Chris Wickham for discussion of this issue.
39
Halsall, Settlement, pp.176–84, 199–211, 219–31.
40
See also above, note 19.
41
Halsall, Settlement, pp.118–42. R . Joffroy, Le Cimetière Mérovingien de Lavoye (Paris,
1974).
Page 236
Ordernr. 061297 P1
ritual focal point, as perhaps it was later by its church, rather than (as yet) defining
itself by delineating its edges: the bounds or borders between itself and its neigh-
bours. That burial rites were a powerful means of shaping communal identity
seems likely. As will be discussed below, they were a focus for competition as well
as the reaffirmation of ties and alliances within the community, and the location
of important economic rites of gift-exchange. It is also noteworthy that sixth-
century cemeteries, or cemetery phases, are larger than seventh-century. Ceme-
teries like Koenigsmacker, Métrich and Puxieux (all destroyed without any
archaeological recording) apparently contained two hundred burials or more, a
figure also estimated for Ennery. The sixth-century phase of Lavoye contained
198 graves, nearly twice as many as in the seventh-century phase (Fig. 10.2), a ratio
also seen at Dieue-sur-Meuse.42
I have previously elaborated an explanation for the furnished burial as a symp-
tom of local competition for power and of the insecurity of local status.43 Analysis
of the grave-goods’ patterns of distribution suggests that the greatest ritual dis-
plays were associated with those people whose deaths caused the greatest rupture
in local social networks: adult males who had died before their sons had estab-
lished themselves, and younger women who were important as the linchpins in
alliances between families. Study of the grave-goods reveals clear local customs
governing the sorts of grave-goods appropriate to particular categories of people:
children generally with few grave-goods and never masculine ones; teenage women
with jewellery; young adult males with weapons; old people with few grave-goods;
and so on. These are most plausibly viewed as communal norms. In sixth-century
cemeteries, local standing was apparently demonstrated, maintained and perhaps
enhanced by the exaggeration of these norms, not by their infringement. Thus,
most of the region’s Adelsgräber or ‘tombes de chef’ are characterized by the inter-
ment of large numbers of the appropriate grave-goods for an individual of that age
and gender. Examples include Lavoye 319, Güdingen 4 and Chaouilley 20c.44 At
42
J. Guillaume, ‘Les nécropoles mérovingiennes de Dieue/Meuse (France)’, Acta Praehistorica
et Archaeologica, 5–6 (1974–75), 211–349.
43
Halsall, Settlement, pp.245–48; developed in G. Halsall, ‘Burial, Ritual and Merovingian
Society’, in The Community, the Family and the Saint: Patterns of Power in Early Medieval Europe,
ed. J. Hill and M. Swan (Turnhout, 1998), pp.325–38; G. Halsall, ‘Burial Writes: Graves, “Texts”
and Time in Early Merovingian Northern Gaul’, in Erinnerungskultur im Bestattungsritual:
Archäologisch-Historisches Forum, ed. J. Jarnut (Darmstadt, 2004), pp.61–74.
44
Lavoye: references at note 41. Güdingen: F. Stein, ‘Les tombes d’un chef franc et de sa
famille à Güdingen: Considérations sur le rôle de l’aristocratie dans l’implantation franque entre
Page 237
Ordernr. 061297 P1
222 Guy Halsall
Ennery, by contrast, there are two burials (graves 70 and 71) that appear to
demonstrate distinction by clear breach of the norms;45 this seems unusual in the
sixth century.
As is well known, sixth-century cemeteries were often organized by rows.
Lavoye’s sixth-century phase is a good example. There were perhaps also norms
about the burial of particular types of people. The sixth-century phase at Lavoye
comprised three distinct groups of burials, suggesting similar internal spatial
arrangements, such as the burials of women mostly being located in the northern
half of the zone. A similar distribution is also notable at Ennery. Child burials are
also found in clusters. This suggests communal organization. It is also interesting
to think about clusters of similar burials. In the past, I have interpreted these as
members of the same family, the usual way of reading ritual similarities. If we see
the development of cemeteries by rows, according to communal organization,
however, perhaps similar burials are those of people of similar social standing (and
age-group and gender) who died close in time. They are part of a ritual dialogue
conducted within terms defined by the villa-community.46 I suspect, furthermore,
that appeal to villa-community custom or norms was one means of keeping
ambitious families in their place, combating their desire to break free of these
norms and mark themselves out as different from their neighbours.
The villa-community might then have been the normal arena for social interac-
tion. It doubtless overlapped and nested within other types of community. It is
probable that kinship networks tended to cross the borders of villa-territories, for
example. Larger-scale geographical territories defining levels of community are more
difficult to perceive. The civitas* probably existed in the north of Gaul, but the
evidence is very unclear. Nevertheless, an analysis of the patterns of lavish burial
in the region suggests that the town of Metz was the focus for some social and
political action. Burials with unusually lavish displays of grave-goods are common
only beyond about a day’s journey from the old civitas-capital. This suggests that
whereas further away from this centre competition for local status was focussed
even more on the burial ritual, close to the city such standing was ascertained by
la Meuse et la Sarre’, Saarbrücker Studien und Materialen zur Altertumskunde, 1 (1992), 117–44.
Chaouilley: J. Voinot, ‘Les fouilles de Chaouilley: Cimetière mérovingien’, Mémoires de la Société
Archéologique Lorraine, 54 (1904), 5–80.
45
Halsall, Settlement, pp.92–93. Alain Simmer subsequently published the site’s data more
thoroughly: La nécropole d’Ennery (Moselle) (Woippy, 1992). This has little effect on my analysis.
46
Halsall, ‘Burial Writes’.
Page 238
Ordernr. 061297 P1
Figure 10.2. The sixth- and seventh-century phases of the cemetery

at Lavoye, ‘La Haie des Vaches’.
Page 239
Ordernr. 061297 P1
224 Guy Halsall
appeal to the secular and ecclesiastical powers based in the city.47 Metz’s ability to
act as such a focus might also have been based upon its function as a major royal
centre. Other civitas-capitals in the region do not appear to have the same effects
upon the distribution patterns of lavish graves.
There were fewer large cemeteries in the sixth century than in the seventh (Fig.
10.3). Although, when looking at the totality of a region’s cemetery data, the evi-
dence is usually poor and partial, the greater numbers of seventh-century cemeteries
cannot result from chance or differential discovery. Sixth-century burials contained
more grave-goods and more distinctive artefact-types (S-Brooches, most bow-brooch
types, bird-brooches, franciscae*, angones* and so on) not found in seventh-century
necropoleis. This means that sixth-century cemeteries are easier than seventh-century
to detect in lacunose written accounts of archaeological discoveries. The contrast
in the density of distribution must be real. This increase in numbers relates not to a
seventh-century expansion of settlement but rather to the relative size of cemeteries
in the two periods. Around 600, the importance of the old cemetery broke down
with the foundation of new graveyards, significantly illustrating the fragmentation
of the earlier villa-communities. Sixth-century cemeteries appear to be used by large
communities, probably drawn from several settlements/residential groups. At
Lavoye, the three groups mentioned above apparently represent the burials of three
different settlements or geographically defined sub-communities, coming together
for burial in the main cemetery. Only one of these groups really continued into the
seventh century (see Fig. 10.2). In association with this development, new small
cemeteries were founded, with fewer grave-goods. In turn this is linked to a switch
to greater investment in above-ground monuments rather than transient grave-
goods. Seventh-century cemeteries were also organized, it seems, by family. Organiza-
tion by rows breaks down and instead we see small clusters of burials of individuals
of all ages and both sexes. Simultaneously, social difference is shown by the breach of
the old norms. This can be seen very clearly at Audun-le-Tiche, where a cluster of
prestigious burials in and around the remains of a Roman fanum complex includes
old women buried with, for that site and period, unusually lavish jewellery.48 Such
items had, in this region during the sixth century, generally been associated with
women between puberty and forty, and usually at the lower end of that age-bracket.
47
Halsall, Settlement, pp.17–18 (continuity of civitas unit); pp.107–08, 260 (distribution of
lavish graves).
48
Halsall, Settlement, pp.158–59. A. Simmer, Le cimetière Mérovingien d’Audun-le-Tiche:
archéologie d’aujourd’hui no.1 Mémoire de l’Association française d’Archéologie Mérovingienne,
2 (Paris, 1988).
Page 240
Ordernr. 061297 P1
Figure 10.3. Distribution of sixth- and seventh-century cemeteries

within the civitas of Metz.
Page 241
Ordernr. 061297 P1
226 Guy Halsall
The explanation for these changes must be sought in greater security of local
status. This led to the weakening of the importance of the old ritual focus and
reduced the need to demonstrate status to a large audience present on the day of the
funeral. Burials began to take place within the villa structures at this time, which
should probably be interpreted as another manifestation of above-ground com-
memoration. A change in the buildings’ function, and possibly their final aban-
donment as any sort of dwelling-place, may have been associated with this. This
increased security and stability of social hierarchy might have helped break up the
community as previously defined (simultaneously, the breakdown of the strength of
the former communities might have allowed this security to emerge) and have given
rise to the situation attested in the later seventh-century charters in other ways.
In the seventh century the increase in data available shows the appearance of
another larger-scale community based around geographical units, above the villa and
below the civitas: that based upon the pagus (there were at least five pagi within the
civitas of Metz). A reference in Fredegar’s Chronicle (dating to the 660s) suggests
that individuals might be defined by their pagus of residence: a man is referred to as
a homo scarponiensis (a man from the district of Scarponne, on the Moselle south of
Metz).49 These pagi seem to replace the civitas as the general unit of administrative
organization — by the Carolingian period at least, they tended to be governed by
counts — although the latter retains importance as a bishopric. The patterns of
documentary survival are doubtless partly to blame, but it remains interesting,
especially given the upsurge in activity in the city from the late sixth century, that
very few Merovingian and early Carolingian charters are issued in Metz, as one might
have expected had the city remained a focus for regional political activity. I suggest
that the appearance of the pagus is part of the same process as saw the break-up of the
old large cemetery-using communities and of the civitas. The more secure status
of dominant families within villae probably led to their establishment of kinship
and other links with like families in other areas, so that the pagus unit became a
more meaningful arena of social action at the level of local and regional elites.
Insiders versus Outsiders
Internal community dynamics might have helped preserve the villa as a unit. So,
too, I suggest, did the relationships between the villa-community and the outside
world. Small arenas of socio-political interaction, local competition and insecurity
49
Fredegar, Chronicle, 4.52, in The Fourth Book of the Chronicle of Fredegar with its
Continuations, ed. and trans. J. M. Wallace-Hadrill (London, 1960), p.43.
Page 242
Ordernr. 061297 P1
of status are not antithetical to large and powerful political units: states or
kingdoms. The Merovingian example makes that very clear. Connection with the
Merovingian kingdom was an effective local political strategy. The families that
had communal pre-eminence in the fourth century might well still have had it at
the end of the sixth but it was, as stated, expensively maintained and resolutely
local. They were unable to break away into a broader socio-political arena. In-
volvement with the Merovingians allowed participation in wider politics, and a
return to something like the fourth-century situation. Adopting ethnic identity
played a part in this, enabling participation in the army and in other types of
politics. This led to another important aspect of this strategy, royal patronage in
the form of office. This not only permitted involvement in wider politics, it was
also a means of securing local power: officials had judicial roles.
Such figures also collected taxes. Here we return to the villa as fiscal unit.
Taxation continued in sixth-century northern Gaul, though its details escape us.
The significance of the vicani dwelling in villae discussed in the Pactus Legis Sali-
cae has been related to fiscal obligation.50 How did it operate in this non-monetary
economy? I have only tentative proposals. I suggest that Merovingian royal
officials and other servants were paid, like late Roman troops and officials, with
drafts on tax revenue and that this is what is meant in sixth-century sources when
a king grants a villa to someone. Here my approach owes much to the work of
Walter Goffart,51 although it must be admitted at the outset that there is little in
the documents to provide a priori support for my thesis. The hypothesis is based
not upon clear empirical support but upon the fact that it makes sense of a
number of issues. This is not to say there are not equally plausible alternatives.
Nevertheless, there is nothing to gainsay the argument either, and some clues add
plausibility to it. One is Roman precedent and in this region, with the military
presence discussed earlier, that might be particularly important. Another is the
Burgundian legal indication (and the implication of Pactus Legis Salicae) that a
villa’s inhabitants shared fiscal liability. Another possible support is the apparent
ease with which villae were granted out and reincorporated into the fisc. In the
Life of Romaric for example, talking about this region, Theuderic II, having killed
50
A. C. Callander Murray, Germanic Kinship Structure: Studies in Law and Society in
Antiquity and the Early Middle Ages (Toronto, 1983), pp.67–87.
51
W. Goffart, ‘From Roman Taxation to Medieval Seigneurie: Three Notes’, Speculum, 47
(1972), 15–87 and 373–94; W. Goffart, Caput and Colonate: Towards a History of Late Roman
Taxation (Toronto, 1974); W. Goffart, ‘Old and New in Merovingian Taxation’, Past and
Present, 96 (1982), 3–21.
Page 243
Ordernr. 061297 P1
228 Guy Halsall
his brother Theudebert II and his supporters in 612, brought the latter’s villae
back into the fisc (villae eorum fisci dicionibus redacte sunt).52 One might also cite
Goffart’s discussion of how different forms or levels of taxation withered,53
though I would read the evidence in a slightly different way. My reading of the
villa is thus that in the later fifth and sixth centuries free inhabitants of these
regions tended to pay dues only to the Merovingian state and its officers. The
precise form which such obligations took might have been related to ethnicity:
Franks being liable for military service but being exempt from some taxation; Ro-
mans paying taxes.54 Rents and other exactions owed, on the basis of land tenure,
to powerful and established aristocrats were not yet common.
Taxation need not have involved very much payment from the locality to the
centre. Perhaps the kings supported themselves essentially from the villae that
they held directly. Various forms of surplus were paid to the king’s designated man
on the spot for the latter’s maintenance. In the region under discussion, to judge
from the later charters, surplus was mainly generated by agriculture, although pig-
farming features significantly. Areas such as the Vosges, where livelihoods might
have been produced by other means, do not appear to have been occupied in the
Merovingian period. The distribution of cemeteries shows that settlement was
restricted to land below 300 m above sea-level. Nevertheless, many could have got
rich on this surplus. There were very wealthy Merovingian royal officials. We
should, however, note that the relationship between a royal officer and those who
lived in the area whence he drew his salary would not make him their lord. This
is important, helping us explain the absence of such relations of domination
(other than those dependent upon royal service) from the earliest Frankish law
code.55 Some tax could have been paid on in bullion. Burials contain many Roman
coins, usually early Roman coins with higher intrinsic worth. As has long been
known,56 balances are important features of very lavish burials in northern Gaul,
52
Vita Romarici Abbatis Habandensis 3, ed. B. Krusch, MGH SRM, 4 (Hannover, 1902),
pp.221–25.
53
Goffart, ‘Old and New in Merovingian Taxation’.
54
G. Halsall, Warfare and Society in the Barbarian West, 450–900 (London, 2003), pp.44–50.
55
A problem brought out by C. J. Wickham, ‘Problems of Comparing Rural Societies in Early
Medieval Western Europe’, Transactions of the Royal Historical Society, 6th series, 2 (1992),
221–46.
56
J. Werner, ‘Fernhandel und Naturalwirtschaft im östlichen Merowingerreich nach archëo-
logischen und numismatischen Zeugnissen’, Bericht der Römisch-Germanischen Kommission, 42
(1961), 307–46.
Page 244
Ordernr. 061297 P1
probably suggesting that a role in controlling ‘weights and measures’ was a signifi-
cant basis of local authority.
People dispatched to collect whatever dues were extracted from these villae
and men sent to support themselves with drafts on taxation would be seen as out-
siders by the villa-community. Here, perhaps, the levying of dues was a means by
which participation in the wider polity helped to define the villa-community. If
the villa was a ‘fiscal asset’, that might have helped identify its inhabitants as a
group. Taxation was, thus, a prize that drew individuals into royal service and out
of their localities, but at the same time, the execution of royal service might have
been a force that maintained the reality and identity of villa-communities.
To explain the changes that took place around 600 we can postulate that the
right to collect royal dues from a designated region was converted into the
outright ownership of the land of that territory.57 In the past I have argued that
this was by royal grant, or grant by those controlling the palace, to buy support in
a period of political crisis during the royal minorities from 575 onwards.58 This
transformation could also have been brought about by force during periods of
weak royal rule. This transformation in the nature of authority is significant in
that it changes the relationship between the aristocrat and the inhabitants of the
villa, transforming the latter into tenants. At this point a surplus-extracting
stratum is increasingly interposed between freeman and state. Hence Ripuarian
Law’s differences from Salic Law in regard to social structure: it refers to freemen
in the service (obsequium) of others, to freemen with their followers (satellites), to
immunities, and to lords answering for their dependants in court.59 Grants of
villae in the seventh-century charters are gifts of outright ownership.
How was the picture presented by the charters brought about? Ownership of
villae could have fragmented through the usual mechanisms of partible inheri-
tance, dower, and so on, as I have suggested in the past. On the other hand,
although this may be plausible and might work in some cases, there is no necessity
57
For analogous processes (sometimes of varying legality), see W. Goffart, Barbarians and
Romans, 418–584: The Techniques of Accommodation (Princeton, 1980), pp.93–102; C. J.
Wickham, ‘European Forests in the Early Middle Ages: Landscape and Land Clearance’, in
Wickham, Land and Power: Studies in Italian and European Social History, 400–1200 (London,
1994), pp.155–99 (pp.162–68).
58
Halsall, Settlement, pp.269–70.
59
Lex Ribvaria 35, 45, 68, ed. F. Beyerle and R . Buchner, MGH Legum Sect. I, 3 (Hannover,
1951). This also has an effect upon the raising of armies, the other key way in which early
medieval kingdoms exacted dues from their subjects: Halsall, Warfare, pp.53–56.
Page 245
Ordernr. 061297 P1
230 Guy Halsall
to adduce it as a blanket explanation. The multiple landholding visible in, for

example, the Wissembourg charters is not necessarily inconsistent with that of
sixth- and seventh-century developments drawn above from the cemetery data.
The extensive ownership of large shares of villae, attested by the charters, is com-
mensurate with the increase in the power and security of local aristocrats in the
seventh century.
What makes my earlier thesis more plausible, however, is the contrast with
royal villae. These differed simply because the fiscal surplus from them had never
been granted away, and had never been turned into outright lordship. Instead they
remained in a direct relationship to the king and so could still be granted away
whole, turning occupiers of the land into tenants, as I suggest had been done from
the late sixth century. These grants are the visible ends of a process that had been
going on for some time.
Conclusions
The thesis set out above presents a model of the evolution from the Roman villa
to the Merovingian through a process of internal dynamics and integration into
wider social and political structures.60 I also suggest how manmade landscape
features provided foci for community identity and indeed for the dynamics that
helped preserve this. The model presented here is of course not the only one
possible. Most of the individual forms of evidence may be read in different ways.
The present hypothesis has some advantages over the alternatives, however. The
first is that it accounts for all of the different forms of evidence. Most other
models concentrate upon one form of evidence and ignore the others. Second, it
accounts for change through time, whereas other explanations have tended to
aggregate data from the whole period between the fifth and ninth centuries, thus
ironing out the situation’s dynamism.
Questions and comparisons nevertheless remain. One of the most pressing is
the comparison of the situation analysed above with the Anglo-Saxon case studies
discussed elsewhere in this volume. It should be clear that, although more plenti-
ful in Gaul, many of the forms of evidence, and the changes therein, as well as the
patterns of data survival, are rather similar in northern Gaul and in lowland
Britain. Might the explanations adduced here be considered in relation to Anglo-
60
For some comments on the later history of the early medieval Frankish villa, see Fouracre,
this volume.
Page 246
Ordernr. 061297 P1
Saxon evidence? The relationships between local community politics and those
of wider polities can be seen in other case studies in this book, notably (if in
slightly different ways) in the case studies of northern Spain. A final advantage of
the model proposed here for northern Gallic developments over its alternatives
is that it pays attention to very particular regional factors. Many other discussions
have a tendency to assume that the situation was everywhere the same in the north
of Gaul (and occasionally elsewhere too). The model proposed in this essay is not
necessarily directly applicable elsewhere, but the dynamics and other factors that
it identifies might profitably be considered in other regional studies.61
61
In addition to general feedback from the other members of the ‘People and Space’ project,
I am grateful to Steve Bassett, Wendy Davies and Chris Wickham for more detailed and very
helpful comment on earlier drafts of this essay.
Page 247
Ordernr. 061297 P1
Page 248
Ordernr. 061297 P1
C OMMUNITY, IDENTITY
AND THE L ATER A NGLO -S AXON T OWN :
THE C ASE OF S OUTHERN E NGLAND
Grenville Astill
W
hen and how did early medieval settlements gain a distinctive urban
identity? And was it derived from the character of the settlement itself
and/or through its relationship to its rural surroundings? The latter
is particularly relevant to our people and places theme: it is important to consider
the effect of urbanization on existing rural communities, and whether, for
example, it changed the way communities organized or defined themselves.
The process of urbanization is seldom considered in such social ways, however,
being more usually approached in economic or political terms. Towns have often
been regarded as concentrations of wealth, which acted as centres of demand that
integrated (and sometimes dominated) surrounding regions. The urban signifi-
cance of a particular place is assessed by the extent to which it stimulated a
coherent and articulated exchange system, indicated, for example, by the degree
of specialized production within its hinterland which made different groups of
producers economically reliant on each other. Alternatively, or concurrently,
towns can be seen as an important means by which royal control was extended
over a kingdom and facilitated the development of medieval states.1
In order to explore the impact of towns on the rural population we have to
consider the social implications of urbanization. We need to return to the urban
sequence, reviewing in particular the information for south-central England between
1
D. Perring, Town and Country in England: Frameworks for Archaeological Research (York,
2002), pp.1–3; C. Wickham, ‘Overview: Production, Distribution and Demand, II’, in The Long
Eighth Century, ed. I. L. Hansen and C. Wickham (Leiden, 2000), pp.345–77 (pp.372–73);
R . Hodges, The Anglo-Saxon Achievement (London, 1989), pp.150–85.
Page 249
Ordernr. 061297 P1
234 Grenville Astill
the ninth and eleventh centuries, with the intention of examining the character
of these ‘urban’ settlements, and what impact they had on their surrounding areas.
An urban sequence common to most of north-west Europe from at least the
ninth century seems to have been established. The documentary, topographic and
archaeological evidence makes it possible to identify settlements with a relatively
diverse occupation structure that developed to serve the needs of aristocratic house-
holds, both secular and religious. In northern France and the Rhineland such set-
tlements, it is suggested, had an initial ‘manorial’ phase of urbanization and trade,
geared to the exchange of produce drawn from their (often far-flung) lands, and
articulated by the estate agents and servants. In the course of the ninth century,
production and exchange intensified, attracting the greater involvement of indepen-
dent traders and the development of regional trade. The role of supraregional and
international trade in this sequence of urbanization is still debated, although recent
attention has been more focussed on agricultural intensification, evidenced mainly
from monastic estates, but also indicated indirectly from increasing evidence of tolls.2
From the later ninth century towns were also developing at a time of agrarian
change, most particularly in terms of the fragmentation of large estates into
smaller agricultural units, which supported a burgeoning lesser nobility. The
changes also affected the peasantry, many of whom, especially in grain-producing
areas, became more tied to the land and burdened with increased obligations.3
Despite these changes in agrarian structure, urbanization continued to be stimu-
lated and dominated by aristocratic production and consumption until at least the
late eleventh/early twelfth century, when there are signs that some of the urban pop-
ulation had achieved sufficient economic and political power to achieve a degree of
independence. This trend had considerably accelerated by the later twelfth century,
when there is extensive evidence for the increased commercialization of the coun-
tryside and a greater involvement of a large part of the population in the market.4
In France, where the scale and the chronology of such economic changes are
debated, it is generally accepted that these took place at a time of political frag-
mentation, when a significant proportion of royal power, especially justice, had
2
A. Verhulst, The Rise of Cities in Northwest Europe (Cambridge, 1999), pp.44–67.
3
For example, R . Fossier, ‘Rural Economy and Country Life’, in The New Cambridge
Medieval History, vol. III: c.900–1024, ed. T. Reuter (Cambridge, 1999), pp.27–63 (pp.42–45,
50–53); R . Faith, The English Peasantry and the Growth of Lordship (Leicester, 1997), pp.153–77.
4
D. Keene, ‘Towns and the Growth of Trade’, in The New Cambridge Medieval History, vol.
IV :
c. 1024–1198, Part I, ed. D. Luscombe and J. Riley-Smith (Cambridge, 2004), pp.47–85
(pp.61–63); R . H. Britnell, The Commercialisation of English Society, 1000–1500, 2nd edn
(Manchester, 1996), pp.79–101.
Page 250
Ordernr. 061297 P1
COMMUNITY, IDENTITY AND THE LATER ANGLO-SAXON TOWN 235
passed into the hands of those lesser aristocrats who controlled their localities.
The focus of their power was often a castle, around which developed a ‘seigneurie
banale’, a district in which the lord (often a former royal official responsible for
public order) exercised power, using his own military might, for his own ends.5
England experienced similar economic changes, although rural restructuring and
urban development is often related to the effects of a growing and centralizing state.
The military reorganization instigated by the kings of Wessex in order to stem the
Viking conquest is seen as initiating the political unification of England. A major
consequence was the creation of a stable environment for economic activity, which
was facilitated by a hierarchical infrastructure that administered law, maintained
order, controlled a strong currency and supervised marketing. In such a situation,
there is a predisposition, especially if a highly efficient or ‘maximum’ view of the
Anglo-Saxon state is taken, to see towns as being initiated and imposed by royal fiat.6
That such diametrically opposed interpretations — French political fragmen-
tation versus English political unification — are proposed for what are essentially
the same economic changes should perhaps make us consider whether it is appro-
priate in this case to view socio-economic change in political terms.
There are also divergent views about the timing of English urbanization.
Firstly, there is a short chronology of urban growth; in this scheme the process of
state creation and military reorganization, usually attributed to Alfred in Wessex,
took place at approximately the same time as increased productivity in the coun-
tryside. The burhs* were commissioned as forts to prevent further Viking advances.
But they were also intended to respond to the greater productivity by becoming
regional economic centres for the exchange and manufacture of commodities.
Such settlements were planned in Wessex by the late ninth century, and extended
to the south-east and east midlands as the kingdom of Wessex expanded during
the tenth century. Some towns, most notably Winchester, provide evidence of
rapid urban growth during this time, in the form of a dense population and diverse
occupational structure, but comparable and contemporary evidence appears to be
extremely limited for the rest of south-central England.7
5
For example, T. N. Bisson, ‘The “Feudal Revolution”’, Past and Present, 142 (1994), 6–42;
D. Barthélemy, ‘Comment 1’, and S. D. White, ‘Comment 2’, Past and Present, 152 (1996),
196–205 and 205–23; T. Reuter, ‘Comment 3’, C. Wickham, ‘Comment 4’, and T. N. Bisson,
‘Reply’, Past and Present, 155 (1997), 177–95, 196–208, and 208–25.
6
For example, Hodges, Anglo-Saxon Achievement, pp.150–85; J. Campbell, ‘The Late Anglo-
Saxon State: A Maximum View’, Proceedings of the British Academy, 87 (1994), 39–65.
7
M. Biddle, ‘Towns’, in The Archaeology of Anglo-Saxon England, ed. D. Wilson (London,
1976), pp.91–150 (pp.120–34); D. Russo, Town Origins and Development in Early England,
c.400–950 A .D . (Westport, CN, 1998), pp.193–231.
Page 251
Ordernr. 061297 P1
The lack of comparable information for other burhs has prompted a second,
and longer, chronology that separates in time the political and military purposes
of the burhs from their economic functions. The burhs were part of a very suc-
cessful military strategy, and continued as a potent symbol and instrument of a
developing state. It is, however, difficult to detect evidence that the burhs had a
simultaneous economic role. The archaeological and the numismatic material
both point to a general increase in economic activity, but only from the later tenth
or early eleventh century; this can be seen in the burhs themselves in terms of
more intense occupation and industrial production. The burhs, then, had an
important military and political role that was retained for perhaps as much as a
century, and only after that did the settlements acquire further urban attributes.8
As both chronologies accept the initial military purpose of the burhs, it follows
that they were superimposed on the landscape according to strategic criteria,
including defensible locations and an even distribution to allow the majority of
the population to take refuge within the defences. Our knowledge of the burhs is
dominated by one document, the Burghal Hidage, which details how each fort
was to be defended. There is a tendency to interpret the document as indicating
an urban framework in Wessex that to a large extent remained unchanged during
the medieval period. This is too static a view. Some forts may have existed before
the Viking attacks and were incorporated into the scheme of defence. Likewise,
there may have been much more potential for change in the military and urban
arrangements during the tenth and eleventh centuries, for example by the com-
missioning of new forts during the raids in the early eleventh century.9
Just as a more dynamic view of military provision may be necessary, we may
also need to take into consideration a similar fluidity in the urban process, from
either the late ninth or late tenth century depending on which chronology —
short or long — is preferred. But both interpretations adopt a primarily economic
explanation of why some burhs became important centres: their success was
dependent on the intensification of both agricultural production and trade. Both
give little consideration to the act of imposing the burhs on Wessex — the mili-
tary advantage would have been all too obvious, and once their success as forts had
8
G. Astill, ‘Towns and Town Hierarchies in Saxon England’, Oxford Journal of Archaeology,
10 (1991), 95–117.
9
D. Hill and A. Rumble, eds, The Defence of Wessex: The Burghal Hidage and Anglo-Saxon
Fortifications (Manchester, 1996); N. P. Brooks, ‘England in the Ninth Century: The Crucible
of Defeat’, Transactions of the Royal Historical Society, 5th series, 29 (1979), 1–20 (pp.17–19);
Biddle, ‘Towns’, pp.140–41; D. Hill, ‘Athelstan’s Urban Reforms’, Anglo-Saxon Studies in
Archaeology and History, 11 (2000), 173–85.
Page 252
Ordernr. 061297 P1
been demonstrated, it is assumed that their value as a different kind of settlement

would have been appreciated.
Wessex Before the Burhs
What effect did the superimposition of the burghal network have on the existing
social and political fabric of the south? Any recognition of the ‘rural community’ is
through the evidence we have for the administrative arrangements, whereby people
assembled at a particular location, drawn from a ‘territory’, for either or both secular
and religious purposes.10 In this sense the ‘community’ is an imposed identity, al-
though it is possible that such units originated from assemblies and their catchment
areas that had been established by common consent at an earlier time and for dif-
ferent purposes. Because the community was an administrative construct, it may only
have comprised those who were important for the maintenance of peace and order
in the regions, and did not therefore constitute the entire rural population. However,
there is also a strong possibility that such administrative constructs could both cut
across or indeed reflect the physical extent of existing communities.
In many regions a coincidence between ecclesiastical boundaries and those for
secular administration — for example those that became later Anglo-Saxon
hundreds* — has been established. In Wessex and the midlands it has been possi-
ble partially to reconstruct a pattern of administrative and religious territories
which were probably at least eighth-century.11 At the core of the territories were
two focal settlement types: the royal estate centre and the minster*. Royal estate
centres are documented as administrative centres (some controlled by reeves) and
collection points for food rents; some existed as hundredal centres by the tenth
century.12 However, before the ninth century there is sparse evidence for such
foci: the most visible were the shire* centres, which in Wessex existed by the mid-
eighth century, but some appear to have declined in political and economic status
10
Summarized for Wessex in B. Yorke, Wessex in the Early Middle Ages (Leicester, 1995),
pp.123–32, and in A. Reynolds, Later Anglo-Saxon England (Stroud, 1999), pp.75–81.
11
P. Hase, ‘The Church in the Wessex Heartlands’, in The Medieval Landscape of Wessex, ed.
M. Aston and C. Lewis, Oxbow Monograph, 46 (Oxford, 1994), pp.47–81 (pp.61–69); P. Hase,
‘The Mother Churches of Hampshire’, in Minsters and Parish Churches: The Local Church in
Transition, 950–1200, ed. J. Blair, Oxford University Committee for Archaeology, monograph
17 (Oxford, 1988), pp.45–66. See Bassett, this volume.
12
P. Sawyer, ‘The Royal Tun in Pre-Conquest England’, in Ideal and Reality in Frankish and
Anglo-Saxon Society, ed. P. Wormald (Oxford, 1983), pp.273–99 (p.281).
Page 253
Ordernr. 061297 P1
by the ninth century (in the sense that their function seems to have been taken
over by other places), such as Somerton, Dorchester and Hamtun.13
More emphasis is now laid on the minsters as local centres and incipient
towns. The evidence tends to concentrate at the beginning and end of the min-
sters’ sequence, from the charter material of the seventh and eighth century, after
which there is often little information until Domesday or a chartered market or
borough in the late twelfth or thirteenth century. The ‘territories’ or parishes* of
minsters continue to be reconstructed, and their extents often coincide with the
boundaries of tenth-century hundreds which, it is argued, probably reflected ear-
lier administrative units. The rural communities identified with the places at the
centre of these ‘territories’ as their religious, economic and, if these places also
contained royal estate centres, administrative foci. Where there was no such coin-
cidence, there was some overlapping of ‘territories’ (for example, where a single
hundred had the same extent as a number of adjacent minster parishes), and so
more places for the communities to assemble.14
Coin finds from minster and royal sites, beginning in the late seventh century
with sceattas*, support the claim that such places functioned as consumption and
collecting centres of agricultural surplus derived from their estates and ‘territories’.
Sceattas are also found at locations which indicate more informal places for sporadic
exchange such as road and river crossings and hill forts.15 In addition, some rural sites
in eastern England have substantial quantities of metalwork as well as coins — the
‘productive sites’. This range of sites could indicate that there was a variety, perhaps
even a hierarchy, of production and exchange centres that may also have supplied
commodities to the coastal and estuarine emporia. Such an economic relationship
may be implied by the extensive distribution of Ipswich ware, a pottery produced
at that emporium. It has also been argued that there is a strong association between
concentrated finds of metalwork and religious sites. Most of this material, however,
cannot be dated later than the mid-ninth century, and there is sparse evidence for
such places existing between the mid-ninth and eleventh centuries.16
13
J. Blair, ‘Small Towns, 600–1270’, in The Cambridge Urban History of Britain, vol. I:
600–1540, ed. D. Palliser (Cambridge, 2000), pp.245–70 (pp.248–49); J. Campbell, ‘Power and
Authority’, in ibid., pp.51–78 (p.52).
14
Blair, ‘Small Towns’, pp.250–53; Bassett, this volume.
15
D. M. Metcalf, ‘Monetary Circulation in Southern England in the First Half of the Eighth
Century’, in Sceattas in England and the Continent, ed. D. Hill and D. M. Metcalf (Oxford,
1984), pp.27–71.
16
J. Moreland, ‘The Significance of Production in Eighth-Century England’, in Long Eighth
Century, ed. Hansen and Wickham, pp.69–104 (pp.87–98); B. Palmer, ‘The Hinterlands of
Page 254
Ordernr. 061297 P1
A significant proportion of minsters have consistent topographic locations

near roads and on promontories overlooking rivers, often confluences. Ovoid or
circular elements in settlement plans are often interpreted as indicating minster
enclosures, as at Bampton (Oxfordshire).17 A limited amount of archaeological
fieldwork has been done on a small number of minster sites. Bicester parish
church, for example, is thought to have been a minster. A section of a ditched
inner enclosure was located and burials were recovered to the north of the church,
and interpreted as part of a large cemetery extending from the church. To the east
the church overlooks the River Bure. The later, twelfth- or thirteenth-century,
town is located on the other, eastern bank of the river, and the two areas were
probably linked by a ford or causeway. Excavations to the south of the medieval
market place, in Bury End, revealed five substantial timber buildings, dated
between the later tenth and twelfth centuries. These are interpreted as part of a
‘proto-town’ that was developing at a short distance from an older monastic nu-
cleus across the river. As the buildings were on a different location and orientation
from the later medieval town, the settlement was probably moved and replanned
with burgage plots and a market place in the twelfth century.18
The Bicester sequence is reminiscent of the more extensively excavated Steyning,
where there was a minster from at least the ninth century. The area to the south
and west of the church was occupied during the eleventh century with low-density,
scattered, timber buildings and rubbish pits, with an enclosed area to the east, a
possible high-status residence. Unequivocal urban characteristics are difficult to
identify here, and it is likely that such settlements only experienced more intense
activity when markets were held. The Domesday entry hints at this with Steyning’s
‘eight burgesses living as villeins’. It is possible the eleventh-century phase of settle-
ment may be related to the time when some minsters had been disendowed — the
holdings of the church in the Domesday entry seem small. We may be seeing a
secularization of the urban process, which might explain the relatively late creation
Three Southern English Emporia: Some Common Themes’, in Markets in Early Medieval
Europe: Trading and ‘Productive’ Sites, 650–850, ed. T. Pestell and K. Ulmschneider (Maccles-
field, 2003), pp.48–60; K. Ulmschneider, ‘Settlement, Economy and the “Productive” Site in
Anglo-Saxon Lincolnshire, AD 650–780’, Medieval Archaeology, 44 (2000), 53–79.
17
Blair, ‘Small Towns’, pp.250–55. However, dating of such features is difficult and secular
sites also had ovoid enclosures, as at Trowbridge: A. Graham and S. Davies, Excavations in
Trowbridge, Wiltshire, 1977 and 1986–88 (Salisbury, 1993).
18
J. Blair, ‘Anglo-Saxon Bicester: The Minster and the Town’, Oxoniensia, 67 (2002),
133–40.
Page 255
Ordernr. 061297 P1
of a mint in the 1020s.19 Such secularized places may have been chosen as mints
as part of a royal policy to recognize or control small towns — indicated in that
category of place mentioned in Edgar’s laws that needed twelve witnesses for trans-
actions (the same as for hundreds), rather than the thirty-six expected for large
towns.20 Places such as Steyning were important religious and economic centres
for their regions, often shown in Domesday and confirmed later by a replanning
of the settlement in the twelfth and thirteenth centuries (see Fig. 11.1).
We are thus dealing with small-scale urbanization, similar to the initial phases
of north European town development mentioned at the beginning. It is also
‘small-scale’ in the sense that these places were related to religious and/or adminis-
trative territories that were smaller than the shire, many of which may later have
been renamed as hundreds from the late tenth century. The sparse archaeological
evidence shows an intensification of settlement centring on the eleventh century,
perhaps related to the disendowment of some minsters and an increasing royal
interest. A major task for the future will be to discover the character of these
settlements in the ninth and tenth centuries.
The Burhs
By the later ninth century this network of small-scale centres was augmented by
the burhs, the ‘planned scheme of national defence’ recorded in the Burghal
Hidage (c.914–18). The locations of most burhs have been identified, although
few have been excavated extensively, and dating remains problematic.21
Burhs were primarily military constructions strategically located, and sited to
offer shelter for most of the population: nowhere was further than 30 km from a
burh. The burhs are customarily classified as towns or forts, the latter being those
sites which had an entirely military and perhaps short-lived purpose. That some forts
had characteristics similar to planned towns of the twelfth and thirteenth cen-
turies encouraged the view that they were also created as towns. Size and a regular
19
M. Gardiner and C. Greatorex, ‘Archaeological Excavations in Steyning, 1992–95: Further
Evidence for the Evolution of a Late Saxon Small Town’, Sussex Archaeological Collections, 135
(1997), 143–71. Similar but more limited evidence has come from Milborne Port, the site of a
minster and royal estate centre: A. Smith, ‘Medieval Archaeological Features at Church Street,
Milborne Port’, Somerset Archaeology and Natural History, 146 (2003), 37–46.
20
Yorke, Wessex, pp.236–38; Die Gesetze der Angelsachsen, ed. F. Liebermann (Halle, 1903),
p.210.
21
Biddle, ‘Towns’, p.124.
Page 256
Figure 11.1. Places mentioned in the text.
Page 257
Ordernr. 061297 P1
street system are such characteristics. The sites selected as burhs included walled
Roman towns and existing small-scale central places. Some were regularly planned
on open ground; others had their streets laid to accommodate the topography,
usually on promontories.22 A high proportion were founded on royal land, which
initially may have facilitated the victualling of the garrison by earmarking food
rents. A variation apparent in the initial scheme was to locate the burh on a royal
estate, but at a distance from the estate centre. This may just reflect the strate-
gically unsuitable location of some estate centres, but it should remind us that,
while the concentration of administrative, military and religious functions in one
place was becoming common, it was by no means invariable and such functions
could be dispersed among nearby settlements. Somerset has the most examples,
such as the minster and shire centre at Somerton and the fort at Langport, or the
minster and royal estate centre at Cheddar and the fort at Axbridge.23
In such defensive schemes the fort may have been the element most susceptible
to change, and revised military plans or a new threat could cause the relocation of
garrisons. The renewed Viking attacks in the early eleventh century led to the re-
commissioning of the hill forts at South Cadbury and Old Sarum.24 In some cases,
for example in Devon, royal confiscation and exchange of land was necessary in
order to build new forts such as Oldaport to control the county’s south coast and
its major estuaries. A similar process of royally enforced land acquisition has been
reconstructed as the prelude to Edward the Elder’s burh-building programme in
the east midlands.25 The fluidity in the burghal network, or evidence for its rear-
rangement, has been claimed in those cases where some forts were apparently
abandoned for another location, such as Halwell for Totnes, Pilton for Barnstaple,
and perhaps the Avebury and Chisbury forts for Marlborough, but the evidence
for any relationship between these pairs of sites has yet to be presented.26
22
Biddle, ‘Towns’, pp.124–37.
23
Also Williton the estate centre and Watchet the fort: M. Aston, ‘The Towns of Somerset’,
in Anglo-Saxon Towns in Southern England, ed. J. Haslam (Chichester, 1984), pp.167–201
(pp.188, 173–74). Examples also occurred in Wiltshire, e.g. Bedwyn and Chisbury (fort):
J. Haslam, ‘The Towns of Wiltshire’, in Anglo-Saxon Towns in Southern England, ed. Haslam,
pp.87–147 (pp.94–102).
24
Biddle, ‘Towns’, pp.140–41.
25
P. Rainbird, ‘Oldaport and the Anglo-Saxon Defence of Devon’, Proceedings of the Devon
Archaeological Society, 56 (1988), 153–64; R . Fleming, ‘Monastic Lands and England’s Defence
in the Viking Age’, English Historical Review, 100 (1985), 247–65.
26
Compare T. Slater, ‘The South-West of England’, in Cambridge Urban History of Britain,
vol. I, ed. Palliser, pp.583–607 (pp.590–91).
Page 258
Ordernr. 061297 P1
Changing military needs, and in some cases the separation of military, religious
and administrative functions, may have meant that this royal burghal initiative
did not, or was not intended to, fit so easily into the older network of regional
centres. There may be no great surprise at this conclusion if the initiative is re-
garded as essentially military, as long as we think of the Burghal Hidage as merely
one record of military organization at one particular time, and that that organi-
zation could change dramatically according to later military needs.
However, if the burhs were intended from the outset to be towns, then their
economic success would depend on a close integration with the networks of trade
that had already been established. We need, therefore, to try to assess the
economic character of the burhs, and how that changed between the late ninth
and the eleventh centuries.
The Economic Character of the Burhs
Winchester is exceptional in the amount of historical and archaeological attention

it has received: the range and the quality of the information demonstrates that by the
late ninth century it had urban characteristics. A major seat of royal and episcopal
authority, it could claim to be a capital until the eleventh century. Much of the
economic activity may well have been geared to serving the needs of elite groups
and had at its heart craft production and the distribution of agricultural surplus
from the estates of religious communities and secular lords. It is generally thought
that the intramural areas became densely populated, and that suburbs began to
develop from the late ninth century, when there is also evidence of diverse occu-
pations. However, some take the view that this increased activity may still have
been mainly associated with, and dependent upon, the churches and palaces, and
that Winchester did not become a large town until ‘well into the tenth century’.27
It is more difficult to assess the character of the other Wessex burhs, but first
it is worth recalling the evidence for royal intention and coercion. From the early
tenth century the laws show a clear concern to supervise trade, to judge from the
emphasis on witnessing transactions and preventing theft. Edward the Elder tried
to concentrate all trade into ‘ports’, but Athelstan’s, Edgar’s and Cnut’s laws seem
to acknowledge that other places existed beyond direct royal control for small-
27
M. Biddle and D. Keene, ‘General Survey and Conclusions’, in Winchester in the Early Mid-
dle Ages, ed. M. Biddle (Oxford, 1976), pp.449–508 (pp.452–54); D. Hinton, ‘The Large Towns,
600–1300’, in Cambridge Urban History of Britain, vol. I, ed. Palliser, pp.217–43 (p.226).
Page 259
Ordernr. 061297 P1
scale trading and allowed them to continue. It was stipulated that borough courts
were also to be held three times a year that should have ensured that suitors came
to town. But, whereas the burhs outside Wessex were also the shire centres, and
therefore the shire and borough courts were held in the same place, there was no
such coincidence in the south; there was therefore potentially less cause for those
representatives of the rural population to acknowledge that the burhs had a major
place in their lives.28
One of the major reasons obliging people to visit burhs would have been to obtain
coin, presumably in exchange for commodities they had produced, in order to pay
taxes. From the 920s there is a steady growth in the number of mints that were
founded in burhs, and this reflects the royal determination to control the coinage,
and to supervise and protect the moneyers and their bullion. Viewed in the light
of the legislation regarding trade, it is further evidence of an attempt to extend
royal control over the local economy. The ability successfully to exercise this level
of control is dependent on how effective the Anglo-Saxon state had become.29
Whether the legislation reflected a high level of trading might be checked by
using coin distributions. The number of coins in circulation during the later ninth
and most of the tenth century was limited. The loss rate of coins in this period was
considerably less than that of sceattas in the eighth century, and this would argue
for a reduced level of trading during most of the tenth century. The eighth-
century level of coin loss was not regained in the south of England until the last
quarter of the tenth century at the earliest, perhaps not until the twelfth century
(Fig. 11.2). Such a reduced level of trading is confirmed to a certain extent by the
remarkable dearth of imported pottery from most inland sites, and even ports,
including London. Rhenish and French pottery reappeared in London and other
ports in the eleventh century.30
28
Discussed in H. Loyn, ‘Towns in Late Anglo-Saxon England: The Evidence and Some
Possible Lines of Enquiry’, in England before the Conquest, ed. P. Clemoes and K. Hughes
(Cambridge 1971), pp.115–28; P. Wormald, The Making of English Law: King Alfred to the
Twelfth Century (Oxford, 1999), pp.289–90, 319–20.
29
T. Reuter, ‘The Making of England and Germany, 850–1050: Points of Comparison and
Difference’, in Medieval Europeans, ed. A. Smyth (Basingstoke, 1998), pp.53–70.
30
D. Hinton, ‘Coins and Commercial Centres in Anglo-Saxon England’, in Anglo-Saxon
Monetary History, ed. M. Blackburn (Leicester, 1986), pp.11–26 (p.18); M. Blackburn,
‘“Productive” Sites and the Pattern of Coin Loss in England, 600–1180’, in Markets in Early
Medieval Europe, ed. Pestell and Ulmschneider, pp.20–36 (pp.31–32); A. Vince, Saxon London:
An Archaeological Investigation (London, 1990), pp.106–08.
Page 260
Ordernr. 061297 P1
Figure 11.2. Single-coin finds from southern England (after Blackburn,

‘“Productive” Sites and the Pattern of Coin Loss in England’, p.32).
Coins and imported pottery, however, may only indicate interregional and inter-
national trade, and it is possible that the burhs had become established as local,
intraregional exchange centres. Local pottery is not sufficiently distinctive, nor
produced in burhs sufficiently often at this time, to be used to estimate the extent or
intensity of local trade. The best guide available is to look at the burhs themselves
and see whether the range and character of occupation is similar to that found at
Winchester. The relatively few excavations that have taken place in the intramural
areas of burhs have shown little evidence for intensive occupation, illustrated most
clearly by the extensive empty, unoccupied, areas in Cricklade and Christchurch.
Similarly, there is no clear sign of residential or industrial occupation — even the
frontages lining the cardinal streets of burhs such as Wallingford remained empty.31
The tenth century, then, appears to be the lost century for the burhs of Wessex.
The success at repulsing the Vikings meant that there was less need of the burhs for
their primary, defensive, purpose and they had yet to fulfil the urban aspirations
which may also have been intended when they were constructed in the late ninth
century. The burhs did not possess the necessary attributes for substantial urban
growth, nor did their location and function coincide sufficiently with the existing old
pattern of centres for assembly and trading to benefit from its continued existence.
31
Excavation results from individual burhs are summarized in Astill, ‘Towns and Town
Hierarchies’, pp.104–08.
Page 261
Ordernr. 061297 P1
The burhs may have been perceived by the majority of the rural population as not
being any more economically differentiated than the older central places, different
only in symbolizing the coercion and control of the growing state.
The ambivalent status of burhs may also be illustrated by the revision of the
dioceses in Wessex in 909. Three new bishops’ seats were created, none within
burhs: Wells and Crediton, where minsters were promoted to cathedrals, and the
same is probably true of Ramsbury. Similarly the male reformed monasteries were
located elsewhere, with the exception of Winchester.32 It seems that royal inten-
tion to keep bishops and burhs apart in Wessex overrode any economic or political
advantage that could have been gained by having episcopal households in towns,
as was to happen soon after the Norman Conquest.
If we are questioning whether burhs fulfilled their economic, and specifically
urban, aspirations, we should go on to ask who were these settlements for and what
happened in them. The burhs’ military purpose required a large aristocratic involve-
ment, specifically in terms of leading, providing and provisioning the garrison. The
association with the royal effort was underpinned by grants of holdings in the burhs
to thegns and higher members of the aristocracy. Documented for Winchester from
the ninth century, the grant-giving is reflected in the Domesday entries for other
burhs, and is often thought to indicate an earlier arrangement associated with the
burghal defensive measures. Indeed some think that the extent of aristocratic interest
in burhs was under-recorded.33 The holdings, ‘urban manors’, were large plots with
buildings in the centre, as reconstructed for tenth-century Oxford (Fig. 11.3) and
Winchester. Occupation of this type might also explain the proliferation of urban
churches as many could have originated as private chapels* for noble households.
Thegns also commonly acted as royal agents in towns, especially as moneyers and
as borough reeves, and it seems that a higher proportion of those thegns with
urban connections or obligations continued to hold office after the Norman
Conquest compared to their rural counterparts.34
32
It is also perhaps significant that the practice of using relics ‘to create a sense of identity for
each town’ in west Mercia was not followed in Wessex: J. Barrow, ‘Churches, Education and
Literacy in Towns, 600–1300’, in Cambridge Urban History of Britain, vol. I, ed. Palliser,
pp.127–52 (p.130).
33
Biddle and Keene, ‘General Survey and Conclusions’, p.457; J. Blair, Anglo-Saxon Ox-
fordshire (Stroud, 1994), p.117; N. Baker and R . Holt, ‘The City of Worcester in the Tenth
Century’, in St Oswald of Worcester: Life and Influence, ed. N. Brooks and C. Cubitt (Leicester,
1996), pp.129–46 (p.140).
34
Blair, Anglo-Saxon Oxfordshire, pp.152–56, 163; R . Fleming, ‘Rural Elites and Urban
Communities in Late-Saxon England’, Past and Present, 141 (1993), 3–37; J. Campbell, The
Anglo-Saxon State (London, 2000), pp.xxv, 191–92.
Page 262
Ordernr. 061297 P1
Figure 11.3. The north-west part of Oxford in the tenth to twelfth centuries
(after Blair, Anglo-Saxon Oxfordshire, p.156).
Page 263
Ordernr. 061297 P1
Aristocratic interest in towns also has an archaeological dimension. The loss

rate of coins in burhs, for example, is most similar to that from ‘thegnly’ rural
residences, suggesting that until the late tenth century burhs were barely differ-
entiated from rural estate centres in terms of trading. A further similarity between
these locations can be seen in industrial activities, especially metalworking. The
‘thegnly’ sites such as Faccombe Netherton in Hampshire have evidence for
metalworking, particularly whitesmithing using gold, silver enamel and glass —
similar activities to that from Winchester and Oxford, in areas interpreted as
urban manors. In both rural and urban locations the scale and character of the
metalworking indicates a craftsman working to satisfy the demands of a patron
rather than mass production for a wider clientele.35
Domesday has been used to demonstrate the character and scale of aristocratic
involvement in towns, but it can also be interpreted as suggesting that by the later
eleventh century this association was diminishing. By 1086 some thegns had
started to lease their urban properties, and others were taking the initiative and
founding their own small markets, as at Newbury, Crewkerne, Great Faringdon,
Yeovil and Okehampton.36
Both trends could indicate the changing character of towns. From the late
tenth century more people were living in burhs: street frontages were colonized,
often with buildings that have timber-lined cellars — the first specifically urban
form of medieval architecture for at least six hundred years — at Wallingford,
Oxford (Fig. 11.3), Chichester and Exeter. Potteries and industries based on rural
surplus such as tanning, bone-working and wood-working became established,
apparently for the first time, in the late tenth to early eleventh century. Such
evidence, and the coins, show that the burhs were becoming towns in the sense
that they were gaining a variety of occupations and a dense population, and were
interacting with the surrounding area in a much more intense and reciprocal way.
The large towns also had a mixed population, with increasingly diverse social and
ethnic groupings, indicated by the dedications and number of the parish churches.
Diverse and wide-ranging contacts also resumed, to judge from the reappearance
of imported pottery, particularly from northern France and the Rhineland.37
35
Hinton, ‘Coins and Commercial Centres’, pp.18–22, 23–24; J. Fairbrother, Faccombe
Netherton: Excavations of a Saxon and Medieval Manorial Complex (London, 1990), pp.244–72.
36
Baker and Holt, ‘City of Worcester’, p.140; Fleming, ‘Rural Elites’, pp.16–17; Slater,
‘South-West of England’, p.594.
37
The details of individual towns are summarized in Astill, ‘Towns and Town Hierarchies’,
pp.103–08.
Page 264
Ordernr. 061297 P1
Suburban development was a further sign of urban growth, spreading along the
major access routes. The building of gateway chapels, now usually dated to the
eleventh century, may indicate a growing self-consciousness among the towns-
people. The eleventh-century rebuilding of the defences at Wareham, Walling-
ford, Cricklade and Christchurch, for example, are usually seen as a response to
the renewed Scandinavian raids and settlement. However, the refurbished
defences, for the first time in stone, could be a further indication that the burhs
were becoming consciously and economically differentiated from the countryside
in general and the small-scale centres in particular.38
The urban momentum established by 1000 is usually portrayed as continuing
throughout the later eleventh and twelfth centuries, and beyond into the thir-
teenth, when the earliest surviving borough records confirm the situation. The
exception was the period immediately following the Norman Conquest when
there is a dramatic, but short-lived, dislocation associated with the destruction of
urban sectors in order to make way for the new castles and relocated cathedrals.
It was short-lived, and was offset by the advantage of having to supply a greater
range of high-status households than before, now that religious and political
institutions had been brought together in towns for the first time.39
However, some aspects of the sparse archaeological record indicate that the post-
Conquest dislocation could also have initiated a process of de-urbanization that
lasted for a large part of the twelfth century. Firstly, it seems that the industrial basis
of the urban economy was not necessarily that secure. Pottery and metalworking,
established and becoming identified as specifically urban activities by the eleventh
century, are not easily found in twelfth-century towns; they appear to have declined
or even disappeared during that century. Just as production is reduced, so apparently
is consumption, for there are fewer twelfth-century portable finds, such as base-metal
jewellery, than previously and the same has been noted for bone items such as combs.
The relative scarcity of mass-produced material, some of it dependent on raw mate-
rials drawn from the countryside, indicates that the century-old relationship between
towns and their hinterlands may not have been that well established, and that the
attraction of the town as a marketing, administrative and entertainment centre was
not necessarily that strong. The twelfth-century town may still have been sustained
economically by acting primarily as a supplier for its institutional residents.40
38
Barrow, ‘Churches’, pp.137–38; G. Astill, ‘General Survey 600–1300’, in Cambridge Urban
History of Britain, vol. I, ed. Palliser, pp.27–49 (p.41).
39
Astill, ‘General Survey’, p.44.
40
Astill, ‘General Survey’, pp.45–46; Hinton, ‘Large Towns’, p.235.
Page 265
Ordernr. 061297 P1
Apart from the substantial buildings occupied by these high-status households,

twelfth-century structures are remarkably elusive in most towns. While due
allowance has to be made for the difficulties of survival, it is possible that this
chronological horizon of buildings was more ephemeral and consisted of earth-fast
timber structures, similar to their rural counterparts, and perhaps providing another
sign of a de-urbanization process. The timber-lined cellared buildings did not
continue to be built in the twelfth century, leaving a hiatus until the construction
of stone cellared and vaulted buildings later in the century.41
The state also had a deleterious impact on the towns. William I imposed a mas-
sive increase of tax on moneyers, from £200–250 to £750–1000, while the
numbers who were exempt from taxation rose considerably, particularly the king’s
followers and Norman townspeople. The effect was to increase the fiscal burden
on an urban society that was already heavily taxed, and contributed to the stultifi-
cation of the economy. Silver shortages between c.1050 and 1160 also contributed
to the potential economic stagnation, and the low loss-rate of coins in towns
during this period would indicate less intense levels of general trading. The
number of mints was progressively reduced from sixty-five in 1083–86 to twenty
in the 1130s to nine in 1189. In some later tenth-century towns the archaeology
shows that moneyers were also engaged in general metalworking, and the con-
nection might help explain this later general decline in urban industry.42
Urban Hinterlands
We need to return to an issue raised at the beginning: whether urban identity can
be established through a settlement’s relationship with its region. It has already
been approached obliquely in the sense that urban economic vitality is often to be
seen in the kind of interaction the settlement had with its hinterland, and it is
clear that the relationship was by no means stable between the late ninth and elev-
enth centuries. Most discussions about hinterlands are based on the assumption
that they were commensurate with the size or complexity of the towns, and that,
once established, the town-hinterland relationship developed in an evolutionary
way.43 The sequence suggested here, however, does not confirm these hypotheses;
rather, for a considerable time the burhs and their occupants were unsure of their
status because the relationship with their region had not been settled, and it was
41
J. Grenville, Medieval Housing (Leicester, 1997), pp.175–81.
42
P. Stafford, Unification and Conquest (London, 1989), p.216; Hinton, ‘Large Towns’, p.235.
43
Perring, Town and Country in England, pp.1–3.
Page 266
Ordernr. 061297 P1
by no means certain that the burhs functioned as significant administrative,

religious or economic centres during the tenth century.
The eleventh century was a critical time, when the potential for the development
of an urban identity was being realized, but was that identity sufficiently established
to survive a de-urbanization phase in the twelfth century? Urban distinctiveness and
town-country interaction may well have been dependent on the extent of continued
royal supervision or intervention. The military origins, the fact that moneyers and
other royal agents were located in places which had separate, borough, courts should
indicate that the burhs were treated in a special way, but not perhaps invariably.
Stafford has reconstructed, mainly from Domesday, the royal estates which supplied
food rents for maintaining the royal court — the farm of one night. She argues that
at the time of the Confessor the rents in Somerset, Dorset, Wiltshire and probably
Hampshire were so large that they were still intended to provide for the whole
court. The groups of estates identified for this purpose were not just agriculturally
productive, but also provided additional income from the hundredal courts and
the burhs which lay within those estates; that is, all the income that was available
from the area was to be used. In these cases any separate identity that the burhs or
hundreds had was ignored, and both were treated as merely a source of income.44
Domesday is also commonly used to reconstruct the hinterlands of towns,
mainly from the arrangements by which urban holdings were attached to rural
manors — the ‘urban fields’. The extent and number of the rural manors that had
interests in any one town is taken as a guide to the status of that town, and thus
some impression of an urban hierarchy can be gained to complement that gained
from numismatic data. Winchester, Cricklade, Oxford and Wallingford had
many dependencies that were spread as far as 40–50 km away, and in the case of
the latter two over several county boundaries. At the other extreme there were
Reading, Calne and Christchurch with one or two attached rural manors within
a radius of 6 km or less: the latter pattern is more typical of the west Wessex towns
(Fig. 11.4).45 Another indication of a town’s influence over its surroundings could
be gained from the few surviving guild documents that record the distance over
which the guilds attracted people. Members of the Exeter guild lived up to 25 km
away from the city, while the Bedwyn guild in the early tenth century made
provision for collecting a member for burial in the burh from up to 80 km away.46
44
P. Stafford, ‘The “Farm of One Night” and the Organization of King Edward’s Estates in
Domesday’, Economic History Review, 33 (1980), 491–502.
45
H. Darby, Domesday England (Cambridge, 1977), pp.309–13.
46
G. Rosser, ‘The Anglo-Saxon Gilds’, in Minsters and Parish Churches, ed. Blair, pp.31–34.
Page 267
Page 268
252
Ordernr. 061297 P2
Figure 11.4. ‘Urban fields’ in southern England (after Darby, Domesday England, pp.311–13).
Grenville Astill
Ordernr. 061297 P1
The most plentiful source for characterizing hinterlands are the coins. After
Edgar’s reform of 973, the coinage was standardized, every issue had the names of
the moneyer and the mint, and on average the coinage was replaced every six years.
We therefore have evidence with good chronological and locational control.
However, coin distributions reflect at least two major functions of coinage —
taxation and trade — and so cannot be treated in the same way as, for example,
pottery; for this reason numismatists correctly warn that coins cannot be used to
write economic history. Coins have, however, been used to produce a ranking of
mints, expressed as a percentage of total known moneyers, which has also been
taken as a guide to the urban hierarchy of late tenth- and eleventh-century
England. It is skewed towards eastern England which reflects the employment of
a greater number of moneyers at the ports, where foreign coin was converted into
legal tender. That there is some economic reality to the importance of the east is
reflected in the distribution of single-find coins, with a marked westward decline
in numbers, indicating that monetization decreased the further west one went.47
The size of the currency was far in excess of what was needed for exchange:
sixty mints were provided whereas perhaps only twenty were necessary, an indica-
tion that coins were not just used for trading but for taxation and as a store of
wealth. Between 973 and 1086 ten mints accounted for two-thirds of the coins,
with London producing 20% — this statistic alone indicates the difficulty of
establishing regional patterns. However, there were varying weight standards
operating within England, within groups of shires. Metcalf has identified seven
regional circulations based on weight, and these have been confirmed by alloy
analysis, showing that the distinctiveness of the localities remained even though
large amounts of coin must have been taken out of the region as tax.48 That towns
and their surrounding countryside had a similar proportion of coins originating
from the same range of mints could be taken as evidence of a significant interac-
tion within particular regions. Fractions — assumed to be used for small-scale
transactions — are not found in greater abundance in the towns. Wessex is
unusual as a region because it has a greater number of mints compared to its share
of national output (8–10%). Many mints are represented by a small number of
47
R . Dolley and D. M. Metcalf, ‘The Reform of the English Coinage under Edgar’, in Anglo-
Saxon Coins, ed. R . Dolley (London, 1961), pp.136–68; D. Hill, An Atlas of Anglo-Saxon England
(Oxford, 1981), pp.130–31; D. M. Metcalf, An Atlas of Anglo-Saxon and Norman Coin Finds,
c.973–1086 (London, 1998), p.34.
48
Stafford, Unification and Conquest, p.142; Metcalf, Atlas of Anglo-Saxon and Norman Coin
Finds, pp.19, 46.
Page 269
Ordernr. 061297 P1
coins, and so many may have operated sporadically. The small mints were serviced
by moneyers from the main town who travelled around the region. Exeter had up
to fifteen subsidiary mints in Devon and Somerset, yet it seems that only five
mints were used in any one issue.49 Coin distributions cannot be used for recon-
structing individual urban hinterlands because a substantial number of coins were
withdrawn from a region in the process of taxation and therefore cannot reflect
the economic potential of individual towns. There is then potential for refining
regional coin-using practices rather than those for a particular place.
A similar conclusion has to be drawn from the region’s pottery. We know that
some towns, such as Exeter and Winchester, had resident potters and the products
were well-produced, but the distribution is quite limited and bears little relationship
to, for example, their ‘urban fields’ which suggests a level of non-economic control
was being applied to these products. Conversely, the coarsewares were very widely
distributed throughout Wessex: one type was used from Christchurch in the west to
Chichester in the east. It may be that sub-groups have yet to be differentiated within
the common coarsewares which would allow a more local view of distribution, but
at the moment the ceramic data can give few insights into urban hinterlands.50
To recapitulate, it is difficult to isolate obvious signs of an urban identity
among the larger towns in the south between the late ninth and eleventh cen-
turies. This may be partly explained by a ‘long chronology’ of urbanization where
the development of clearly urban characteristics was drawn out and faltering. But
there were other contributory factors: the fluidity in the arrangements of the
‘burghal system’ — where urban functions were not only sometimes divided
amongst two or three settlements, but also illustrated by the amount of rationali-
zation, leading to the abandonment of some sites and the promotion or creation
of others — may have been underestimated. And lastly we need to consider the
extent to which the burhs of the late ninth century were successfully accommo-
dated within the older, pre-burh, pattern of trading and assembly places, increas-
ingly recognized as having been based on royal centres and minsters. In particular,
we may need to take more seriously the possibility that this older pattern, despite
royal efforts to the contrary, continued to determine the social and economic
relationships of the majority of the population.
49
Metcalf, Atlas of Anglo-Saxon and Norman Coin Finds, pp.83, 246–48.
50
M. Mellor, Medieval Ceramic Studies in England (London 1994), pp.72–76; D. H. Brown,
‘Pottery and Late Saxon Southampton’, Proceedings of the Hampshire Field Club and Archaeo-
logical Society, 50 (1995), 127–52 (pp.133, 143–45); D. H. Brown, Pottery in Medieval
Southampton c.1066–1510 (York, 2002), pp.8, 127.
Page 270
Ordernr. 061297 P1
M ARMOUTIER: F AMILIA VERSUS F AMILY .

T HE R ELATIONS BETWEEN M ONASTERY AND S ERFS
IN E LEVENTH -C ENTURY N ORTH -W EST F RANCE
Paul Fouracre
I
n 1077 one Waleran who held a villa* at Nanteuil in the Dunois in north-
western France gave all his servi* and ancillae* from the villa to the monastery
of Marmoutier, which lay on the north bank of the River Loire, just upstream
of the town of Tours (Fig. 12.1). In the charter of donation Waleran made it clear
that these serfs were now dependants of the monastery:
Whatever rights I formerly had over them, or over their possessions, let them now belong
to St Martin and the monks of Marmoutier. If any of the children of these serfs, male or
female, should move to another place, be it near, be it far, and go and live in another villa,
vicus, castellum or civitas, having been bound by the obligation of servitude, let it be held
for them there.1
The terms of this grant tell us much. We learn that people could be given away,
just as land or livestock or rights could be given away. We are told that the chil-
dren of the serfs will also belong to Marmoutier, and that this burden they could
not escape, even if they moved away from Nanteuil. Their servitude was thus
personal and perpetual. Finally, we see four types of places in which such people
might dwell. We cannot be sure exactly what is meant by the terms villa, vicus*,
castellum*, and civitas*, but they probably refer to an estate such as the one at
Nanteuil, a village, a fortification and a town, such as that of Tours. The im-
pression we are given is of a landscape in which there was a variety of settlements,
1
Liber de Servis Majoris Monasterii, ed. Ch. L. Grandmaison, Publications de la Société
archéologique de Touraine, 16 (Tours, 1864), hereafter SM, cited by document number. The
documents appended by Grandmaison to the original cartulary are cited as SMA. Here: SMA 31.
Page 271
Ordernr. 061297 P1
256 Paul Fouracre
Figure 12.1. Principal places mentioned in charters relating to Marmoutier.
and of a servile population that was liable to move around, but also of a servitude
that was immutable. What we also know from Marmoutier’s charters is that the
monastery had been building up its influence in this area since the beginning of
the eleventh century, and that it had faced stiff opposition from some of the local
notables there. We know too that serfs such as Waleran’s former dependants
might bitterly resent being tied forever to the monastery, especially if they did
move away and prosper. We are thus dealing with a landscape in which there was
a great deal of wrangling over both people and space.
This essay will focus on the relationship between Marmoutier and its serfs as
revealed largely through one late eleventh-century cartulary*, ‘The Book of the
Serfs of Marmoutier’ (Liber de Servis Majoris Monasterii), and it will discuss the
landscape in the terms just outlined. It will seek to explain why Marmoutier’s hold
over people, rights and land was so often contested, and to understand how the
monastery became a supralocal lord. We can then think about what effect the
monastery might have had on the formation of local communities. But first it
must be admitted that while it is possible to write a coherent history of Marmou-
tier’s expansion in the eleventh century, many features in the local background
will remain hidden from us. For we know much more about religion, war and
power than we do about settlement and local communities. It is therefore not
Page 272
Ordernr. 061297 P1
MARMOUTIER: FAMILIA VERSUS FAMILY 257
surprising that peasant society in north-west France in this period has often been
viewed through the prism of lordship. In an important paper published in 1980,
R. I. Moore made a bold effort to tackle this ‘top down’ model in the context of
the so-called ‘Gregorian reform’ of the church.2 Moore pointed up the community
dynamic that fuelled ecclesiastical reform from below. In particular he emphasized
the importance of the parish* church as the focus of the newly formed commu-
nity. Villagers, he argued, wanted decent parish priests unencumbered by family
responsibilities, and who might act as mediators between village and lord. It
seems, however, that Moore’s community, clustered around the church, was
largely imagined, or at least that he was taking his cue from numerous French
studies that worked backwards from later medieval or modern villages and
deduced that they had been formed as a result of the breakup of the great estates
which had characterized the Carolingian order.3 Halsall in his essay shows how
such estates emerged in north-eastern France as a ranked society was transformed
into communities dominated by a landholding aristocracy. The latter formed an
elite that had acquired land and power by working in partnership with the Mero-
vingian rulers at the turn of the sixth and seventh centuries. In the eighth and
ninth centuries the pattern of widespread landholding would be consolidated and
an archipelago of great estates came to form the basis of Carolingian order. Hence
the notion that when that order broke down, one consequence would be a trans-
formation of settlement patterns and estate structures. When Carolingian
authority decayed, new spaces were supposedly opened up to be filled by now
independent local lords. In this view, lords had encouraged the regrouping of
settlements to their own advantage: a nucleated settlement was easier to exploit
than the kind of scattered settlements evident in the early Middle Ages. The
villagers sought a protector in the parish priest, as a response, Moore thought, to
pressures from above. Nice though this argument is, it tells nothing about how
villages were actually formed, nor about how they were organized.
2
R . I. Moore, ‘Family, Community and Cult on the Eve of the Gregorian Reform’,
Transactions of the Royal Historical Society, 5th series, 30 (1980), 49–69.
3
See, for instance, J. Chapelot and R . Fossier, The Village and House in the Middle Ages
(London, 1980), pp.129–35. The authors admit that it is not possible to see how villages were
formed, but that their formation should be deduced from the breakup of the ‘large farming units
of the early Middle Ages’. G. Fourquin, ‘Les temps de croissance’, in Histoire de la France rurale,
vol. I: Des origines à 1340, ed. G. Duby (Paris, 1975), pp.370–547 charts the growth of lordship,
castles, population, ploughs, etc., but can find nothing to say about the growth of villages, apart
from new settlements on cleared land (p.431), hence the need to deduce their formation.
Page 273
Ordernr. 061297 P1
258 Paul Fouracre
The assumed contrast between the great estate of the Carolingian period and
the nucleated settlement of the central Middle Ages is one that cannot easily be
tested. According to Chapelot and Fossier the fragmentation of the great estate
can be seen in documents from the eleventh and twelfth centuries in which small
units of property are visible, the assumption being that these are subdivisions of
once larger units.4 But as we shall see, one of the chief difficulties in assessing
change over the period is the changing nature of the written record. Records are
much more numerous, fuller and more detailed from the eleventh century on-
wards, ushering in what R. E. Barton has recently termed a ‘new narrativity’.5 Are
we therefore seeing new things in these records, or old things described for the
first time, or at least in unprecedented detail?6 One reason for suspecting that sub-
division is not new is that we can indeed detect it from the seventh century on,
whenever the record is full enough. For example, the villa of Echternach, which
lay midway between the rivers Meuse and Moselle, perfectly fits Halsall’s model
of a large estate formed in the early Merovingian period to support a service
aristocracy. Nevertheless, by the year 706, when a more detailed description of
Echternach first appears in a document, the estate had already been divided into
several portiones.7 Finally, Moores’s account also fails ultimately to escape the ‘top
down’ model: lords still determine changes on the ground.
The idea that central authority broke down at the end of the tenth century,
leading to a so-called ‘Feudal Revolution’, has recently come under strong criti-
cism for drawing too sharp a contrast between supposed Carolingian order and
alleged ‘feudal’ disorder, and for an overly deterministic reading of what little
evidence there is.8 Here is not the place to follow the rather interesting debate
4
Chapelot and Fossier, Village and House, p.135.
5
R . E. Barton, Lordship in the County of Maine, c. 890–1160 (Woodbridge, 2004), pp.12–15.
6
It was with these questions that D. Barthélemy launched his first challenge to the notion of
social and political transformation around the year 1000: D. Barthélemy, ‘La mutation féodale,
a-t-elle eu lieu? (Note critique)’, Annales ESC, 47 (1992), 767–77.
7
P. Fouracre, The Age of Charles Martel (London, 2000), pp.45–47.
8
The essence of the criticism and response to it are in D. Barthélemy, S. D. White, T. Reuter
and C. Wickham, ‘Debate: The Feudal Revolution I–IV’, Past and Present, 152 (1996), 196–223,
and 155 (1997), 177–208. The debate was sparked by T. Bisson, ‘The “Feudal Revolution”’, Past
and Present, 142 (1994), 6–42. For an excellent bibliography of the ‘mutation issues’, see S. D.
White, ‘Tenth-Century Courts at Mâcon and the Perils of Structuralist History: Re-reading Bur-
gundian Judicial Institutions’, in Conflict in Medieval Europe: Changing Perspectives on Medieval
Culture, ed. W. C. Brown and P. Górecki (Aldershot, 2003), pp.37–68 (pp.37–38, n. 2).
Page 274
Ordernr. 061297 P1
about how, or even whether, society was transformed in this period, but it is im-
portant to emphasize the way in which communities have been imagined in the
light of a model of transformation which is itself now under fire. The root of the
problem where north-western French communities are concerned is that elev-
enth-century charters from this area (and from most other areas too) tell us very
little about space. They have much to say about people, frequently naming serfs,
for instance, but they give little indication of where these people actually lived.9
Charters from England, by contrast, have much to say about space in the form of
boundary clauses, but very little about the people who lived within those boun-
daries. This means that, for different reasons, on both sides of the Channel it is
hard to match people to space, but in England, as both Basset and Reynolds and
Langlands demonstrate elsewhere in this volume, it is easier to describe how boun-
daries changed as new units were carved out of older and larger ones. Incidentally,
it is worth pointing out that much of the unflattering contrast English historians
routinely make between England and the Continent in the late Saxon period
comes from the difference in the nature of the charters. Where we have people we
have noise, dispute, and conflict — a nearly dysfunctional society in English
eyes.10 In England, where we have boundaries we may hear of disputes between the
powerful, but the community is silent, and the landscape looks very orderly.
Marmoutier’s Rise in the Eleventh Century
The monastery of Marmoutier was founded in c.371 by St Martin as a retreat

from nearby Tours where he was bishop. Marmoutier acquired enormous prestige
from its founder, who became one of Francia’s premier saints. Although the
institution seems to have waned in importance and influence after the Merovin-
gian period, the association with St Martin would remain a key asset throughout
9
Continental charters for earlier periods also generally give little information about
boundaries, but there are significant exceptions to this rule, such as charters from ninth-century
Brittany: see W. Davies, Small Worlds: The Village Community in Early Medieval Brittany
(London, 1988).
10
In the words of the late Patrick Wormald, ‘There is [. . .] reason to believe that the English
scene was quite unlike that in Professor White’s Touraine: until, that is, conquest superimposed
the French scene upon it’. P. Wormald, ‘Giving God and King their Due: Conflict and its Regu-
lation in the Early English State’, in P. Wormald, Legal Culture in the Early Medieval West
(London, 1999), pp.333–57, at p.341. Stephen White is a leading scholar on conflict in eleventh-
century Touraine.
Page 275
Ordernr. 061297 P1
260 Paul Fouracre
the Middle Ages. At the end of the tenth century Marmoutier was effectively
refounded as a Cluniac house. The two earliest accounts of the monastery’s re-
foundation position themselves in a world of decline and destruction, in this case
making the Vikings the cause of Marmoutier’s decline in the tenth century.11
Though composed in the later eleventh and early twelfth centuries respectively,
these narratives are firmly in the Cluniac tradition of writing about reform and
renewal. It is important to recognize the strength of this tradition because it
directly affects views of the landscape: destruction and emptiness are emphasized
in order to explain the need for renewal and to point up the achievement of
reform. Sometimes this rhetoric has been taken too literally, as in Southern’s mas-
terpiece The Making of the Middle Ages, which opens with an Angevin landscape
still to recover from Viking destruction, a land that had been so severely depopu-
lated and so long left fallow that the very fertility of its soil had been restored.12 It
is a view that has contributed much to the sense that the late tenth and early
eleventh centuries ushered in a new age. One wonders whether the rhetoric of
depopulation in the northern meseta and sub-meseta, with which Julio Escalona
and Iñaki Martín Viso contend in this volume, is part of this same tradition.13
Familiar phrases of emptiness and renewal are certainly found in Catalan charters
further to the east in the later tenth century.14 They are also to be found in
English narratives of the so-called ‘tenth-century reformation’.15 In all of these
areas one could argue that reform was actually triggered by a growth in population
and prosperity rather than as a response to decline and destruction. On the other
hand, this religiously inspired rhetoric of reform is important evidence for a
contemporary sense of change and novelty across Europe.
11
S. Farmer, Communities of St Martin: Legend and Ritual in Medieval Tours (Ithaca and
London, 1991), pp.78–116.
12
R . W. Southern, The Making of the Middle Ages (London, 1967), p.17.
13
In addition to their contributions here, see also J. Escalona Monge, ‘Comunidades, terri-
torios y poder condal en la Castilla del Duero in el siglo X ’, Studia Historica: Historia Medieval,
18–19 (2000–01), 85–120, and S. Castellanos and I. Martín Viso, ‘The Local Articulation of
Central Power in the North of the Iberian Penninsula (500–1000)’, Early Medieval Europe, 13
(2005), 1–42 (pp.30–31).
14
J. Jarrett, ‘Power Over Past and Future: Abbess Emma and the Nunnery of Sant Joan de
les Abadesses’, Early Medieval Europe, 12 (2003), 229–56.
15
For caution against taking this rhetoric too literally in the English context, see C. Cubitt,
‘The Tenth-Century Benedictine Reform’, Early Medieval Europe, 6 (1997), 77–94.
Page 276
Ordernr. 061297 P1
The two narratives of Marmoutier’s reform privilege the comital houses of

Anjou and Blois respectively, the earlier highlighting the importance of Angevin
help, the later, the aid from Blois.16 Marmoutier lay on the River Loire roughly
midway between Anjou and Blois (Fig. 12.1) and was in effect able to play the two
houses off against each other. Conversely, the monastery’s opponents were able
to appeal to Anjou and Blois for help against Marmoutier. In addition to the
counts, there were also viscounts who feature as both supporters and opponents
in our documents, as well as local milites*, and, of course, bishops. Marmoutier
was not the only monastery trying to establish itself along the Loire and its tribu-
taries. There was also St Aubin at Angers, another late tenth-century refounda-
tion, and La Trinité at Vendôme, consecrated in 1040. To the north lay St
Vincent of Le Mans, and to the south, near the fortification of Nouâtre where
Marmoutier controlled several serf families, lay the monastery of St Mary of
Noyers, founded c.1031. None of these foundations was as big or powerful as Mar-
moutier, but they share the same profile in terms of receiving gifts and expanding
their lands and resources. Foundation, or refoundation, was followed by a rush of
donations that reached a plateau around the middle of the eleventh century and
began to decline in the later part of the century. By the middle of the twelfth
century, these monasteries were receiving relatively few donations.17 By the mid-
eleventh century, the Loire valley and its tributaries was seemingly congested with
competing institutions and families. This was what R. I. Moore has termed a time
of ‘hectic generosity’ as local families, mostly of lesser noble status such as the
milites, tried to consolidate their standing by giving to monasteries and receiving
from them spiritual privileges.18 Equally hectic was what Moore calls the ‘counter-
attack’ by nobles ‘competing desperately amongst themselves for land’. This,
explains Moore, is why at a time when ecclesiastical institutions were the bene-
factors of donations on an unprecedented scale, they were also locked in intermi-
nable disputes about resources, and often faced violence from the very families
they counted amongst their clients and donors. An example would be Marmou-
tier’s struggle to keep control of a church at Naveil, just outside of Vendôme.
Sometime between 1009 and 1027 Marmoutier bought the rights of this church
with the permission of its bishop, the renowned Fulbert of Chartres. This meant
16
Farmer, Communities of St Martin, pp.78–116.
17
S. D. White, Custom, Kinship, and Gifts to Saints: The ‘Laudatio Parentum’ in Western
France 1050–1150 (Chapel Hill, 1988), pp.20–25. The Appendix, pp.212–34, has excellent
tables breaking down the business of each monastery by twenty-five-year period.
18
R . I. Moore, The First European Revolution c. 970–1215 (Oxford, 2000), p.82.
Page 277
Ordernr. 061297 P1
262 Paul Fouracre
buying out the bishop’s tenant and his sub-tenant, and the latter’s sons, and their
mother. The initial arrangement was then challenged no less than eleven times up
to the year 1072.19 Other parties claiming an interest were bought out, and as time
passed more and more of the descendants of the original tenants demanded
money as compensation for what they regarded as a loss of inheritance. Marmou-
tier seems time and again to have been prepared to pay rather than fight. By 1072
it had paid out over 340 solidi* and seven pounds of silver, in addition to the
original ‘not small’ purchase price. For one group to give and for another to con-
test the gift is certainly not new, but some novelty lies in the scale and intensity
of the giving and the taking, and also in the detail in which this traffic was
recorded. Making and keeping detailed records was a way of managing change.
Marmoutier itself produced three cartularies in the later eleventh and early
twelfth centuries. These are cartularies for the Dunois and the Vendômois and the
‘Book of Serfs’.20 In addition historians have reconstructed cartularies for the
Blesois, La Perche and for the Tours area itself.
Monastic Familia and Serf Families
The ‘Book of Serfs’ consists of 127 notices all concerned with serfs, although the
other cartularies also contain material dealing with serfs.21 It was put together
around the year 1070, but the bulk of the notices come from the time of Abbots
Albert and Bartholomew, thus 1032–64, the time when giving to the monastery
was at its height. The impetus behind collecting these notices may have been a
kind of stock-taking after the death of Bartholomew and, probably, after the death
of a prior called Odo who was Marmoutier’s chief agent in the Vendômois.22
19
The business of the church at Naveil is the subject of the first twelve documents of the
monastery’s cartulary for the Vendômois: Cartulaire de Marmoutier pour le Vendômois, ed. M. de
Trémault (Paris, 1893).
20
Cartulaire de Marmoutier pour le Dunois, ed. M. Emille Mabille (Chateaudun, 1874);
Cartulary for the Vendômois, above note 19; the ‘Book of Serfs’, above note 1. On the
composition of the various cartularies, see D. Barthélemy, ‘Note sur les cartulaires de Marmoutier
(Touraine) au XIe siècle’, in Les Cartulaires, ed. O. Guyotjeannin, L. Morelle and L. Parisse (Paris,
1992), pp.247–59.
21
These are the documents Grandmaison appended to his edition of the ‘Book of Serfs’.
22
The penultimate notice about the church at Naveil from 1072 refers to Odo’s death three
years earlier, and notes his importance in settling previous disputes: Cartulaire de Marmoutier
pour le Vendômois, ed. de Trémault, no. 11.
Page 278
Ordernr. 061297 P1
What the notices do in effect is to record the servile status of individuals who had
either recently become serfs of the monastery or who had contested their status. The
monastery went to great lengths to get recalcitrant serfs to acknowledge their status
because this determined the status of all their descendants, and so the ‘Book of
Serfs’ worked to safeguard future resources that could be expected to grow. What
the serfs actually did in terms of service, or owed as tribute, was not recorded.
Many of the notices recorded ‘auto-deditions’, this being when people volun-
tarily gave themselves to the monastery to become its serfs.23 They publicly
acknowledged their newly chosen status by placing four denarii* (pennies) on
their heads, and it was declared that they and all their descendants would remain
serfs of Marmoutier, forever. Grandmaison, the editor of the ‘Book of Serfs’,
believed that the situation in the countryside of the Touraine must have been
truly terrible for free people to enter into servitude in this way. Here he was
mindful of a well-known early medieval formula* from Tours in which desperate
people commended themselves to a lord (becoming his serfs) in return for guaran-
tees of basic sustenance and shelter.24 Dominique Barthélemy, however, has read
the notices in a much more positive light.25 He argued that those giving them-
selves into servitude were hardly desperate peasants, and most of them were not
peasants at all. Becoming a serf of the monastery might actually be a good career
move, thought Barthélemy, when the newly enserfed received lands on favourable
terms, or filled important posts within the monastic familia*, such as bailiff or
cellarer. In addition, such people would receive spiritual privileges in return for
giving themselves. Barthélemy termed them ‘ministerial serfs’, by which he meant
to indicate that servitude was not a bar to social and economic improvement, and
the difference between freedom and servitude could be illusory. Put this way, one
can see the attractions of being part of Marmoutier’s familia. As we shall see,
membership of this community, which conferred the (unspecified) benefits of its
societas on its more privileged members, was sufficiently valuable to be offered as
inducement to people to settle their disputes with the monastery. Such privileged
people might include serfs as well as the nobility. In this sense the familia can be
23
Examples of auto-deditions: SM 3, 17–20, 25–27, 29, 30, 45, 46.
24
Liber de Servis, ed. Grandmaison, pp.xii–xvii. The Tours ‘commendation formula’ is For-
mulae Turonenses, no. 47, in Formulae Merowingici et Karolini Aevi, ed. K. Zeumer, MGH Leges
in quarto, Sectio 5 (Hanover, 1886), p.158.
25
D. Barthélemy, La mutation de l’an mil, a-t-elle eu lieu? Servage et chevalerie dans la France
des X e et XIe siècles (Paris, 1997), pp.57–91. This is the most extensive modern study of the Mar-
moutier material.
Page 279
Ordernr. 061297 P1
264 Paul Fouracre
seen as a voluntary community that grew whenever and wherever people joined
it. We might see here a realization of that claim to infinite space which, elsewhere
in this volume, Sennis explains was part of the monastic imagination. In practice,
however, Marmoutier’s community also had an involuntary aspect because many
serfs were included in it against their will. This element can be seen to include
some of the descendants of those who had given themselves into serfdom, for, as
we have already noted, the gift of self (auto-dedition) always included future
progeny. The concept of a community that could flow into any person or any
space here comes up against the very practical difficulties of subordinating needs
of the serf family to the interests of its own familia.
It is indeed true that many of the Marmoutier serfs look relatively prosperous,
and they certainly were of sufficient importance for their voices to be recorded in
our notices. But they remained serfs, which means that they faced restrictions on
their movement, on marriage, and on inheritance. It was clear that these restric-
tions would apply to all their descendants, and it is reasonable to think that the
more prosperous families became, the more they would resent restrictions on
marriage and inheritance. Although families might gain immediate advantages by
becoming serfs, subsequent generations of the family might equally wish to cast
off their servile status. In 1069 we hear of one Conversa who had been married to
a certain Otbert. Otbert had become a serf when he took up the tenancy of Mar-
moutier land and became one of the monastery’s ‘mayors’. This meant that his
wife was now a serf, as were all the children born to them. After Otbert’s death,
and after the death of all the children except one son and one daughter, Conversa
petitioned the abbot to free her daughter so that she could marry a free man. Abbot
Bartholomew agreed, but only on condition that the family leave the land, that
the son and his descendants remain serfs, and that should the girl fail to find a free
husband, she would have to return to servitude.26 Very few of these notices record
manumissions, and where they do occur, it is other people who free the serfs.27
With the exception of Conversa’s daughter, Marmoutier itself frees no one.
There were many ways into serfdom, but no way out without permission of the
owner.28 According to the evidence of the ‘Book of Serfs’, Marmoutier acquired
its serfs in four main ways: by gift, through auto-dedition, by marriage (when a
Marmoutier serf married, the spouse also became a Marmoutier serf) and by
punishment, or in lieu of debts. It was the last two means, marriage and punish-
26
SM 76.
27
Examples: SM 13, 51, 52.
28
For an example of a serf buying freedom from an owner, see SMA 18.
Page 280
Ordernr. 061297 P1
ment, which most often led to disputes in which people refused to accept their
servile status. As we have already noted, the monastery was prepared to go to great
lengths to win these disputes because what was at stake was future generations of
serfs, and for the same reason the serfs sometimes fought on for generations. For
example, some time before 1067 one Otbert, a shepherd and a free man married
an ancilla of Marmoutier.29 He now became the monastery’s serf, or, according to
another notice, he was made a serf because he had burned down one of the
monastery’s barns. When the ancilla died, Otbert married a free woman called
Plectrude, and he began to assert that he was also free. Prior Odo got him to admit
his servile status, and this meant that Plectrude was now a serf too. ‘Not wishing
to leave her man’, she made a public acknowledgment of her status. The couple
had a son, Vitalis. After Otbert was dead, Plectrude claimed that Vitalis was free
because he had been born before Otbert had acknowledged his servile status. She
agreed to go to the ordeal of hot iron to prove Vitalis’s right, but when the iron
was heated she lost her nerve, and the case. But Vitalis fought on for the next
generation, only finally yielding in 1096, when Pope Urban II came to bless
Marmoutier’s new basilica.30
Marriage between free and unfree people was naturally a source of friction as
the monastery sought to own the children of such marriages. Marriages between
serfs of different lords were likewise problematic since both lords might claim the
children of these unions and could, as we shall see, insist that children be divided
between the different lords. Cross-lordship marriages also made serf inheritance
complex. A particularly good example of the frictions over inheritance is provided
by a notice of a dispute that ended around 1084.31 It concerned three serfs,
Hilduin, Guy and Herbert, who were serfs of the count of Anjou but whose
maternal grandparents had been serfs of Marmoutier. They had cousins on the
Marmoutier side. On the death of the grandparents, all the cousins wished to
divide up their inheritance, which included houses and vineyards, so that we must
imagine that these people were relatively well off. Marmoutier judged that
Hilduin, Guy and Herbert had no right to a share in this property. The serfs then
turned to their lord, Fulk, count of Anjou, ‘believing that they could get, through
29
SM 108.
30
SM 127. Plectrude’s refusal of the ordeal is discussed by S. D. White, ‘Proposing the Ordeal
and Avoiding It: Strategy and Power in Western French Litigation, 1050–1110’, in Cultures of
Power: Lordship, Status and Process in Twelfth-Century Europe, ed. T. Bisson (Philadelphia, 1995),
pp.89–123.
31
SM 116.
Page 281
Ordernr. 061297 P1
266 Paul Fouracre
the violentia of their lord the count, what they had not been able to get in the
placitum* (the hearing held by Marmoutier)’. Fulk heard their case but judged
against their claim. They then turned to other means, and continued to lobby the
count until there was another hearing before Abbot Bernard. Again their claim
was turned down, and the serfs continued to harass the monastery until Marmou-
tier finally settled by paying Hilduin, Guy and Herbert fifteen pounds of silver to
quit the claim and state that none of their relatives and future descendants would
ever raise it again. The count gave guarantees that if they did return to the claim
he would expel them from his strongholds and fortifications. And this is one of
the examples in which contumacious serfs were also given spiritual benefits (‘our
benefits and society’) as an incentive to settle. The dispute had probably lasted
about twenty years.32 The persistence of both sides in this case is striking, as is the
fact that despite their obvious wealth and influence this family group was unable
to share an inheritance across lordships. It was said to be the ‘law’ (lex) that did
not allow them to do so, and this was a principle so strong that the count could
not overturn it. In a context in which there was fierce competition for resources,
it is not surprising that a law that prevented wealth from being transferred from
one lordship to another should be vigorously maintained.
A second case shows the lengths to which serfs might go to try and prevent
their wealth being divided between different lords, and it further highlights their
vulnerability in matters of inheritance.33 Sometime before 1037 Hilducia, an
ancilla of the viscount of Blois, had married a Marmoutier serf, Ohelm. Ohelm
himself had been given to the monastery by the count and countess of Blois, but
seems to have resented his new position for he was described as ‘always rebellious
and treacherous’. Fearful that Hilducia’s former owners might cause trouble,
Ohelm put up the money to buy her freedom, and a charter of manumission was
drawn up, witnessed by the viscount and his wife. Hilducia was thus temporarily
free, but once she married Ohelm she became a serf of Marmoutier. The couple
then had a son, Ascelin, who seems to have prospered. In a charter of c.1040 the
monks had got Ascelin to swear obedience to them, to stop him turning out like
his father.34 They were particularly concerned that he would marry a free woman
and then make trouble by getting powerful people to support him. Ascelin did in
fact associate with the families of local milites, witnessing several charters from the
32
It seems likely that it began after the death of Abbot Bartholomew in 1064 and it was
concluded in 1084.
33
SMA 18, 27.
34
SMA 7.
Page 282
Ordernr. 061297 P1
Dunois, including a royal praeceptum*.35 After the deaths of Ohelm, Hilducia and
Ascelin, the descendants of Hilducia’s original owners began to fight the mon-
astery for Ascelin’s property. No heir is mentioned. The case turned on the status of
Hilducia at the moment of her marriage to Ohelm. If she had not been free, then
they could claim Ascelin as their serf and have rights to his property. The monks
brought out the charter of manumission, but their opponents refused to look at
it. They then turned to violence, harassing the monks until Marmoutier bought
them out. In the next generation the dispute was reopened, and again violence was
employed to force the monks to settle. Once more the monks bought out their oppo-
nents, and conferred spiritual benefits upon them to seal the settlement. Marmoutier
clearly had the cash to secure its future resources in this way, but its ability and
willingness to pay might help explain why so many people challenged it.
So far we have seen cases in which the serf voice is strong because the serfs had
the wherewithal to fight. People of lesser standing and influence seem to have
been treated more precipitately and may as a result have suffered even more dis-
ruption in the formation of families and communities. One Gandelbert, for
example, denied that he was serf. Prior Odo took him from Vendôme to Marmou-
tier where he was simply imprisoned ‘for a long time’ until he acknowledged his
status.36 In 1087 Marmoutier and a Walter Rimand found that they had servi and
ancillae in common, presumably as a result of intermarriage. They proceeded to
divide up the children of these serfs, splitting up siblings.37 Finally one little girl
was left, not worth assigning because she was still in the cradle and might die
anyway. So the decision on her was postponed. What it actually meant for siblings
to be assigned to different lords in this way is not revealed. But we should imagine
that even if family units were allowed to stay together, the division would affect
future marriage choices and prevent a common inheritance in the next generation.
This case, and many others from the ‘Book of Serfs’, gives one the impression
that Marmoutier and other lords might have control over individual families in
one locality or community in which several lords might have dependants. That
would certainly seem to follow from the gifts of individuals to the monastery, for
presumably some at least of these people continued to live amongst fellows who
had not been given away. Crucial here was the principle of personal dependency,
35
Cartulaire de Marmoutier pour le Dunois, ed. Mabille, 17, 22 (the praeceptum), 48, 49, 132.
36
SM 106.
37
SMA 36. That siblings were split up is indicated by the naming of a single child from a
group of more than one, e.g. de infantibus Rotberti Chiphardi, Adelitia who went to Marmoutier,
meaning, presumably, that the other infantes would have gone to Walter Rimand.
Page 283
Ordernr. 061297 P1
268 Paul Fouracre
so that if people did move, they carried ‘the obligation of servitude’ with them, as
we saw in Waleran’s grant with which this essay opened. Marmoutier had serfs
over a very large area: locations in the ‘Book of Serfs’ stretch from the Limousin
to the borders of Normandy. The key to maintaining control over people across
this great space was to identify them as serfs, to make sure that they acknowledged
their servitude, and to make sure that their children were not lost to the monas-
tery’s familia. For instance, some time between 1032 and 1064 Marmoutier was
given two coliberti*. The monks then identified the two bastard sons of the
coliberti and went on to identify the children of these sons who lived near the
fortification of Nouâtre.38 Listing and naming was all important here, and especial
care seems to have been taken to identify people who lived at greater distance
from the monastery. When the monastery was given two servuli over 150 km away
to the south, near Limoges, the monks recorded the names of the servuli, the
names of all four of their parents, their place of birth, the name of the local castle,
and the name of its castellan. This was so that the descendants of the servuli could
in turn be identified ‘that not just they, but all those born to them in the future
should serve the abbot and brothers of Marmoutier as serfs, for all the days of their
life’.39
Because it had interests in so many different places, Marmoutier could be
described as a ‘supralocal’ lord. It did not, however, have a districtum, an area in
which it exercised rights over all the inhabitants. Nor are there references to it
having its own forces or fortifications. It was not in this sense a ‘banal lord*’. The
picture we get of lordship from the ‘Book of Serfs’ and the other Marmoutier
cartularies is one in which a pack of local notables seems to have worried at every
resource. What we see is a massive space over which the Marmoutier community
was spread, but spread very thinly. This space included a great variety of interest
groups and families across the social spectrum, ranging from serfs, through milites,
to comital and episcopal elites. This plurality of groups exercised a range of
powers, derived from a mixture of ‘official’ authority, spiritual authority, wealth
and force. The variety is reflected in the complex ways in which disputes were
settled, drawing on law, custom, norms, and brute force, and also in the way in
which disputes could be extremely protracted. The cases we have discussed here
were held in different venues: in the chapter house of Marmoutier itself, in the
38
SM 48. How colberti differed from servi is not clear from this material. Coliberti could be
given away, and they laboured under the same perpetual servitude as the servi. See Barthélemy, La
mutation, p.123.
39
SM 75.
Page 284
Ordernr. 061297 P1
Figure 12.2. Priories of Marmoutier established in the eleventh century (after

Gantier, ‘Recherches sur les possessions et prieurés de l’Abbaye de Marmoutier’).
count’s court, before the prior, and in the Ohelm case, in the meadow between the
hall and the tower at Blois. We may see here a topography that matches the
plurality of interest groups.
Marmoutier managed its interests through a network of local cells and agents
that developed into a system of priories (Fig. 12.2).40 Priories came into being
when the monastery had acquired a concentration of land and rights in a given
locality.41 They consisted of a handful of monks and their prior, and served both
as spiritual centres and as points from which resources were collected and con-
trolled. The priories were in theory under the strict control of the abbot, and were
the means by which the monastic familia could spread itself over such a massive
space. Chamars and Lavardin, two places prominent in the ‘Book of Serfs’ seem
40
O. Gantier, ‘Recherches sur les possessions et les prieurés de l’Abbaye de Marmoutier du
e
X au XIIIe siècle’, Revue Mabillon, 53 (1963), 93–110 (maps after p.136).
41
The term prioratus first appears at the end of the eleventh century. Before this there was a
variety of terms (cella, obedientia, locus, domus), or reference was simply made to the monks of St
Martin in a location other than Marmoutier itself; Gantier, ‘Recherches sur les possessions et les
prieurés de l’Abbaye de Marmoutier’, pp.96–99.
Page 285
Ordernr. 061297 P1
270 Paul Fouracre
to have been the first areas in which the monastery’s resources were organized in
this way. By the middle of the thirteenth century a great chain of Marmoutier
priories stretched from southern Aquitaine to England.
It was the prior who dealt with the serfs in a given area. We have seen from the
cases discussed here the importance of Prior Odo in enforcing serfdom in the
Vendôme area. He held initial hearings on questions of status, he carried out the
monastery’s decisions, and he inflicted punishment, including imprisonment. It
was the prior who monitored serf marriages, identified children, and reported on
disputed inheritances. Major disputes were held before the abbot, with the local
prior as a key witness. When serfs publicly acknowledged their status, either at
moments of auto-dedition or when they were forced to do so, this took place in
Marmoutier itself before an audience of the whole community. No doubt the
imposing architecture of the abbey and the size and illustriousness of the com-
munity played a part in making such occasions solemn and binding. But the key
link between the serf and the monastery remained the prior. Relations between
the priories and the monastic centre would therefore directly affect the impact
that Marmoutier had on the lives of its serfs. They would, in turn, determine the
nature of the familia in the local context.
Conclusion: A New World?
Barthélemy’s insistence that there was little new in relations between lords and
serfs in the eleventh century forms part of a wider ‘anti-mutationist’ argument. In
his view, remember, historians have mistaken the changing nature of charters,
with their greater detail, for change itself. One can answer this point by arguing
that the production of more, and more detailed, records is itself likely to have
been a response to change. In the case of Marmoutier, the production of large
numbers of documents both charted the massive inflow of gifts to the monastery
and dealt with the complaints, claims and quarrels that followed in the wake of
this generosity. According to evidence such as the tortuous history of the
monastery’s attempts to control the church at Naveil, competition for resources
was intense. If this picture holds good for relations between lay lords in general,
then the eleventh-century Touraine must have been seething with disputes, many
of which no doubt involved violence.42
42
For an excellent discussion of the place of violence in disputes in this area, see S. D. White,
‘Feuding and Peace-Making in the Touraine around the Year 1100’, Traditio, 42 (1986),
195–263.
Page 286
Ordernr. 061297 P1
It is of course impossible to be categorical about how much was new in the

eleventh century precisely because the changing nature of the record limits com-
parison with earlier periods. A further problem where servitude is concerned is the
persistence of conservative terminology: servus, for example, is the term for slave
in classical Latin as well as for the kinds of serfs we have been meeting, and they
were certainly not slaves.43 Accounts of how one gets from the slave of antiquity
to the serf of the central Middle Ages tend to conform to the positions historians
hold on the wider issue of post-Carolingian transformation (or ‘mutation’). But
the history of servitude is much more complex than the stark binary oppositions
that characterize the transformation debate will allow. Firstly, it is not possible to
speak of slaves or serfs in general without being aware of considerable regional
variation, and secondly we must note that there was differentiation within the
servile class at all periods. It is this variation that made it possible, for example, for
Barthélemy to argue that ‘ministerial serfs’ are visible in the Carolingian period.
He is right. Indeed, we can trace them back before this, to, say, the early seventh-
century ‘Will’ of Bishop Bertramn of Le Mans where they are actually called
ministeriales.44 But the argument must turn not on whether such people existed,
but on whether they were more common in the eleventh century than at earlier
times. One reason for thinking that the latter might be the case is an association
between differentiation and increasing economic activity.
In the ‘Book of Serfs’ we find three basic (and ancient) principles governing
relations with serfs: there is only serf and free, nothing in between; servility was
personal (that is, the person carried the burden with them); serfs could not inherit
outside the lordship. These conservative principles were put into practice in a
landscape in which there was an increasing amount of cash. There was also some
population movement into new settlements and around new fortifications. At the
same time, as we have seen, Marmoutier massively expanded its holdings of land
and people, just as other monasteries were doing so. The serfs we see in our cartu-
lary are nearly all likely to have been acquired by Marmoutier in the two genera-
tions leading up to the ‘Book’s’ composition, at a time when there were unprece-
dented opportunities for serfs to move and to prosper. This is the background to
43
For a discussion of the terminological problems, see H.-W. Goetz, ‘Serfdom and the
Beginnings of a “Seigneurial System” in the Carolingian Period: A Review of the Evidence’, Early
Medieval Europe, 2 (1993), 29–51.
44
Testamentum Bertramni Cenomannis, in Diplomata, Chartae, Epistolae, Leges aliaque
instrumenta ad Res Gallo-Francicas spectantia, ed. J. M. Pardessus, 2 vols (Paris, 1843–49), I,
197–215 (p.208).
Page 287
Ordernr. 061297 P1
272 Paul Fouracre
the very high levels of contestation we see in all the Marmoutier cartularies. There
does seem to be novelty here, in degree if not in kind. By acquiring serfs in many
different localities, by discouraging marriage across lordships and preventing
cross-lordship inheritance, and by insisting on personal dependance, Marmoutier
is likely to have frustrated the consolidation of serf families. As we have seen,
Marmoutier privileged its own familia at the expense of the serf family.
We began by noting how difficult, or even impossible, it is to chart the evolu-
tion of the village in north-west France. It supposedly developed as a result of the
breakup of the large farming units of the early Middle Ages. But as already noted,
these large estates had always been subject to division between heirs. Even the
massive ecclesiastical estates of the early Middle Ages, which were in principle
indivisible, were in practice let out in farming units to clients, sometimes in their
thousands.45 This is of course part of the process known to historians as ‘manori-
alization’, which has been traced back to the seventh century. Most of the farming
units named in the Carolingian polyptychs* of the ninth and tenth centuries
would appear later as villages, and many later medieval villages have names that
were already centuries old. In other words, most places that would become villages
had had people in them for a very long time. What distinguishes the village from
the earlier settlement is the evident level of communal organization.46 This may
yet again be an impression distorted by the increasing amount of information
about local organization that becomes available to us from the twelfth century
onwards. On the other hand, our evidence from Marmoutier suggests that the
monastery’s dealings with its serfs would be inimical to the formation of villages
because the monastery sought to control individuals or families in communities
or settlements in which several lords were likely to have interests. No village
community was likely to form where the monastery insisted on the priority of its
own familia over the interests of individuals and families, and this goes for other
lords too. The striking image of Marmoutier and Walter Rimand splitting up
families in their own respective interests suggests a very high level of intervention
in the serf community. Conditions in the twelfth and thirteenth centuries look
more propitious for the growth of villages. Firstly, serf movement into the familia
45
The monastery of St Wandrille on the lower Seine, for instance, was recorded for the year
787 as possessing 3964 mansi*, 2395 of which were let out to clients: I. Wood, ‘Teutsind, Witlaic
and the History of Merovingian Precaria’, in Property and Power in the Early Middle Ages, ed.
W. Davies and P. Fouracre (Cambridge, 1995), pp.31–52.
46
This is the central theme of L. Genicot’s survey, Rural Communities in the Medieval West
(Baltimore, 1990).
Page 288
tails off.47 Secondly, as priories became more numerous and more widespread,
links between the monastery and the priory became weaker, and Marmoutier’s
centralized control slackened.48 The growing stability of the serf population and
localization of the familia to the priory, as well as falling levels of violence between
competing lords, must all have contributed to a more favourable environment.
This is of course only one aspect of village formation, and there are a host of other
factors — economic, demographic, agronomic, climatic, administrative, etc. — to
be taken into consideration. The evidence from Marmoutier is, however, impor-
tant in showing us how status determined relations people had both with the local
community and with the monastic community. It was only when the latter
intervened less in the former that villages could be organized in the form that we
find them from the central Middle Ages onwards.
47
The last charter of auto-dedition is from 1113/14, and the last gift of a serf is from 1223.
48
Farmer, Communities of St Martin, pp.128–34.
Page 289
Ordernr. 061297 P1
Page 290
Ordernr. 061297 P1
N ARRATING P LACES:
M EMORY AND S PACE IN M EDIEVAL M ONASTERIES
Antonio Sennis
A Foundation
A
round the year 742, after spending three years in Westphalia as a mission-
ary, Sturm — a Bavarian who, as a child, had entered the monastery of
Fritzlar and had been educated by the illustrious Abbot Wigbert — was
summoned by Archbishop Boniface. Boniface instructed him to go around the
forests that surrounded the course of the river Fulda and look for a deserted loca-
tion in which to found a monastery. With the two companions the archbishop
had given him, Sturm left and began to search for a site suitable for a monastic
community. Once they had arrived at a place called Herfeld, the three monks
established a dwelling and started to live there in meditation and contemplation.
After some time, Sturm went back to Boniface and described ‘in detail the loca-
tion, the quality of the soil, the running water, the springs and valleys and
everything else connected with that place’. The archbishop replied to him that,
although he appreciated the fact that they had found a place where to live, he was
afraid to allow them to continue to stay there, because that location was too close
to where the as yet unchristianized Saxons were settled. He therefore suggested
that his disciple look somewhere else, for a place where an ascetic life could be
conducted without danger. Sturm resumed his quest (exploratio in the words of
his biographer) along the course of the Fulda, inspecting ‘the land, the mountains,
the hills, the heights and the valleys’.
More time passed and, wanting to know if any progress had been made, the
archbishop sent for Sturm once again. Receiving him in a secluded area of his
palace, he reassured him that the predestined site of his new abbey had not yet
been revealed, but that it soon would be. In fact when Sturm eventually found the
Page 291
Ordernr. 061297 P1
276 Antonio Sennis
right place, he understood this as divine inspiration and, delighted and filled with
joy, blessed it and carefully marked it. That was the site on which the monastery
of Fulda would be erected (Fig. 13.1).1
The location of a newly established monastery was a very important matter,
even if the choice of site was determined by factors upon which the narratives of
foundation often touch only implicitly.2 In early medieval Italy monastic proper-
ties were, for example, generally defined as areas having the main site as their
centre (terra circa monasterium posita), and in many ninth-century cartularies*
documents were organized according to administrative districts, as is the case of
those produced at Fulda, Mondsee and Wissembourg.3 A community’s inclina-
tion to control and organize the surrounding landscape — which was reflected in
the frequent reference to a monastery as the place (locus) par excellence — was, as
well as having some direct administrative purposes, also justified by the fact that,
in the monastic mentality, secular space had a meaning only insofar as it was
purged by the presence of a monastery. In addition, a monastery was a sacred
place, the point of intersection between the material and the spiritual worlds and,
as such, a place of power.4 It is therefore no surprise that, for example, the
1
P. Engelbert, Die Vita Sturmi des Eigil von Fulda: Literarkritisch-historische Untersuchung
und Edition (Marburg, 1968), pp. 132–43 for these events; pp. 135 and 136 for the two quoted
passages; p. 138 for Sturm as explorator, and his quest exploratio. On this episode, see also
C. Wickham, ‘European Forests in the Early Middle Ages: Landscape and Land Clearance’, in
Wickham, Land and Power: Studies in Italian and European Social History, 400–1200 (London,
1994), pp. 155–99 (pp.156–59).
2
A. Remensnyder, Remembering Kings Past: Monastic Foundation Legends in Medieval
Southern France (Ithaca and London, 1995), p. 23.
3
Ideologically asserted centrality of a monastery with reference to the surrounding landscape:
L. Lagazzi, Segni sulla terra: Determinazione dei confini e percezione dello spazio nell’alto medioevo
(Bologna, 1991), pp. 87–88; A. Remensnyder, ‘Tographies of Memory: Center and Periphery in
High Medieval France’, in Concepts of the Past: Ritual, Memory, Historiography, ed. G. Althoff,
J. Fried and P. J. Geary (Washington and Cambridge, 2002), pp. 193–214. Examples of
cartularies organized according to administrative districts: Urkundenbuch des Klosters Fulda, ed.
E. E. Stengel (Marburg, 1956) and Codex Diplomaticus Fuldensis, ed. E. F. J. Dronke (Cassel,
1850); Das älteste Traditionsbuch des Klosters Mondsee, ed. G. Rath and E. Reiter, Forschungen
zur Geschichte Oberösterreichs, 16 (Linz, 1989); Traditiones Wizenburgenses: Die Urkunden des
Klosters Weissenburg, 661–864, ed. K. Glöckner and A. Doll (Darmstadt, 1979).
4
Monastery as locus by antonomasia: A. Dimier, ‘Le mot locus dans le sense de monastère’,
Revue Mabillon, 62 (1972), 133–54. Foundation of a monastery as an act of cleansing the impure
landscape: G. Penco, ‘Un elemento della mentalità monastica medievale: la concezione dello
spazio’, Benedictina, 35 (1988), 35–71 (p.56); Remensnyder, Remembering Kings Past, p. 48.
Page 292
Ordernr. 061297 P1
NARRATING PLACES 277
Figure 13.1. Main places mentioned in text.
Monasteries as sacred places, by definition, in the territory: G. M. Cantarella, ‘Lo spazio dei
monaci’, Settimane di studio del centro italiano di studi sull’alto medioevo (Uomo e spazio nell’alto
medioevo), 50 (2003), 804–47 (p. 845). Sacred places as sites of power: D. L. Carmichael, ‘Places
of Power: Mescalero Apache Sacred Sites and Sensitive Areas’, in Sacred Sites, Sacred Places, ed.
D. Carmichael, J. Hubert, B. Reeves and A. Schanche (London, 1994), pp.89–98 (p.89).
Page 293
Ordernr. 061297 P1
278 Antonio Sennis
anonymous monk who, at the beginning of the twelfth century, composed the
Miracula Sancti Adalardi stated in his prefatio that the pleasantness of the site and
the richness of the edifices, while being a clear sign of God’s benevolence towards
his abbey, were also sufficient to prove Corbie’s splendid past (Fig. 13.1). He thus
started his narration with a description of the various churches and of the dif-
ferent features of the surrounding landscape.5 And this is also the reason why the
choice — or, better, the discovery — of the site of the abbey was usually narrated
as an almost supernatural exploration, an adventure in which it was actually God
who guided the founder of the community.6 On the other hand, the foundation
of a monastery caused in itself a redefinition of the relations between people and
space. By entering a monastic community, men and women had to relocate
themselves both physically and mentally. It is true that this repositioning did not
completely obliterate the place of origin, and conflicts and divisions within a
monastic community had often, for example, an undeniable (if, at times, slightly
confused) ethnic connotation. However, individuals entering a monastery had to
abandon their place of origin and their old selves, and find new connections in a
largely new social and physical environment.7
5
De miraculis s. Adalhardi, ed. O. Holder-Egger, MGH SS, 15.2 (Hannover, 1888),
pp.859–65 (pp.860–61).
6
On monastic narratives of foundation, see J. Kastner, Historiae fundationum monasteriorum:
Frühformen monasticher Institutionsgeschichtsschreibung im Mittelalter, Münchener Beiträge zur
Mediävistik und Renaissance-Forschung, 18 (Munich, 1978); Remensnyder, Remembering Kings
Past; A. Sennis, ‘Tradizione monastica e racconto delle origini in Italia centrale (secoli XI– XII)’,
Mélanges de l’École française de Rome, 115 (2003), 181–211.
7
Examples of ethnic conflict within a monastic community: Sennis, ‘Tradizione monastica
e racconto delle origini’, pp.207–09, and, for an episode involving Alcuin, Martinus Turonensis,
Vita Alcuini abbatis Turonensis, ed. W. Arndt, MGH SS, 15.1 (Hannover, 1887), ch. 18,
pp.193–94. Solemn vows that each individual entering a monastery had to pronounce: Benedicti
Regula, ed. R. Hanslik, Corpus scriptorum ecclesiasticorum Latinorum, 75 (Vienna, 1960), ch.
58, 17–29. Monastic compulsory threefold abandonment (personal patrimony, family ties and
homeland): G. Constable, ‘The Ceremonies and Symbolism of Entering Religious Life and
Taking the Monastic Habit, from the Fourth to the Twelfth Century’, in Settimane di studio del
centro italiano di studi sull’alto medioevo (Segni e Riti della Chiesa Altomedievale), 33 (1987),
771–834. My use of the category of ethnicity here takes into account the concepts discussed, for
example, in P. Geary, ‘Ethnic Identity as a Situational Construct in the Early Middle Ages’,
Mitteilungen der anthropologischen Gesellschaft in Wien, 113 (1983), 15–26; W. Pohl and
H. Reimitz, eds, Strategies of Distinction: The Construction of Ethnic Communities, 300–800
(Leiden, 1998).
Page 294
Ordernr. 061297 P1
A Prophetic Vision
Around the year 845 Anskar, the first great apostle to the North, was planning a
new journey to Sweden, in order to continue his mission among those people. The
Christianization of mainland Scandinavia had been successful only intermit-
tently: it was some years since a cleric had been preaching in those lands and the
region was at risk of lapsing back into paganism. This is what made both a new
mission highly necessary and Anskar, its main advocate, somewhat distressed and
anxious about it.
One night the saint had a vision. Sitting on a richly decorated throne and dressed
like a prophet, Adalhard, the great but by then deceased abbot of the monastery
of Corbie, where Anskar had been a pupil, appeared to his disciple. The purpose
of the vision was to confirm to Anskar that his journey was going to be blessed by
God’s will and benevolence. In order to reassure Anskar that his mission was des-
tined to success, Adalhard quoted some biblical verses: ‘Hear you islands, and give
ear you peoples from afar [. . .]. I have placed you here as a light for other nations.
You must take my saving power, even until the end of the earth.’ According to his
biographer, these words made Anskar even more certain that he was being sum-
moned by divine command ‘because almost all that country consisted of islands’
and ‘because in the north the end of the world lay in Swedish territory’.8
Ecclesiastical communities had a strong inclination to look at the space around
them in order to find that the perfect correspondence between divine will, natural
laws and social order was materially reaffirmed in it. In his Vita Columbani, written
in the first decades of the seventh century, Jonas of Bobbio recounted, for exam-
ple, how one day Columbanus was considering whether he should undertake the
evangelization of the Wends, a Slavic people. As he was ready to make his vows an
angel appeared to him and drew a sketchy map of the world (mundi compagem
monstravit). ‘Look here,’ this angelic vision said to Columbanus, ‘how the world
remains a desert. Go right and left, wherever you want, in order to eat the fruits
8
Rimbert, Vita Anskarii, ed. W. Trillmich, in Quellen des 9. und 11. Jahrhunderts zur
Geschichte des Hamburgischen Kirche und des Reiches, ed. W. Trillmich and R . Buchner
(Darmstadt, 1961), ch. 25, pp.82–86. References to Anskar’s distress for the fate of the missions
and anxiety for his journey are, respectively, at p.82 and at p.84; the vision is at pp.84–86. The
biblical passage quoted is Isaiah 49. 1. Many of Anskar’s visions are characterized by the under-
taking of a voyage (P. E. Dutton, The Politics of Dreaming in the Carolingian Empire (Lincoln and
London, 1994), pp.51–53). On the theological and literary facets of the Vita Anskarii, see
I. Wood, ‘Christians and Pagans in Ninth-Century Scandinavia’, in The Christianization of
Scandinavia, ed. B. Sawyer, P. Sawyer and I. Wood (Alingsås, 1987), pp.36–67 (pp.39–42).
Page 295
Ordernr. 061297 P1
280 Antonio Sennis
of your labour.’ Hearing these words, and looking at that representation of the
world, the saint understood that it was better if he changed his mind, and stayed
where he was, until the time came for him to undertake his journey to Italy.9
As in Anskar’s dream, in this episode of the life of Columbanus space was used
as the principal medium to convey a divine message, and an understanding of the
sense of the vision occurred without any problem or doubt. This happened
because saints were considered to be blessed by the grace of having the chance to
behold portions of the world usually forbidden to human eyes. In his Commentary
on Scipio’s Dream, a text of pivotal influence in the development of early medieval
monastic culture, Macrobius argued for example that the vacuity of human glory
was demonstrated by the fact that it could only extend to very limited sections of
physical reality, so narrow and mundane when compared with the immensity of
the cosmic space, perceivable only by the divinity.10
This meant that the (mutable) landscape in which monastic communities were
set could be, and actually was, perceived, described, and sometimes imagined as a
forceful representation of an eternal truth. In other words, in the early Middle
Ages, nowhere were those elements so thoroughly combined and structured as in
monasteries — in a system of thought that organized the surrounding world by
placing the community at both its physical and its spiritual centre. In most cases
these communities looked at the physical space around them in order to find their
claims of social supremacy confirmed.
The geography of the site could therefore be described in order to certify this
claimed supremacy. In his account of the foundation of the monastery by
Emperor Louis II, the author of the Chronicle of S. Clemente at Casauria (Fig.
13.1), wrote that the emperor ordered the relics to be put on the back of a donkey.
The animal was then let free to go wherever it wanted; it encountered a river, the
river Pescara, miraculously floated on the water, and stopped in the middle of it,
9
Ionae Vitae sanctorum Columbani, Vedastis, Iohannis, ed. B. Krusch, MGH SRG in usum
scholarum, 37 (Hannover, 1905), I.27, p.217. This reference to the act of drawing mundi
compages somewhat predates any surviving medieval map (see P. Gaultier Dalché, ‘De la glose à
la contemplation: Place et fonction de la carte dans les manuscrits du haut Moyen Âge’, in
Settimane di studio del centro italiano di studi sull’alto medioevo (Testo e immagine nell’alto
medioevo), 41 (1994), 693–764 (p.697)).
10
Macrobius, Commentarium in somnium Scipionis libri duo, ed. L. Scarpa (Padua, 1981). A
similar theme is developed in Gregory of Tours Vita Sancti Venantii (Vitae Patrum, 17.5, ed.
B. Krusch, MGH SRM, 1.2 (Hannover, 1885), p.732). On the influence of Macrobius’s text on
medieval monasticism, see P. Courcelle, ‘La vision cosmique de saint Benoit’, Revue des études
augustiniennes, 13 (1967), 97–117 (pp.110–14).
Page 296
Ordernr. 061297 P1
where there was an island. That was the site of the new abbey (whose full name
was Saint Clement de Casa aurea in Insula Piscarie). The chronicler obviously
wanted to establish a direct link between the story of how the saint’s body was
miraculously found floating by his disciple Phoebus, in spite of the fact that
Aufidianus, the prefect, had ordered him to be drowned in the sea with an old
anchor attached to his neck. The monastery was not built on an island yet that
description was essential because physical and spiritual spaces needed to be made
to agree in order to produce a coherent memorialized identity.11
The Return of Some Relics
Some time during the year 893, the monks from the abbey of St Vaast in Arras,
nowadays in north-eastern France (Pas-de-Calais), addressed a request to the bishop
and the citizens of Beauvais: they wanted the remains of their patron back. Twelve
years before, the community had in fact been obliged to abandon its site due to the
attack of a band of Vikings, whose raids had been tormenting the northern shores
of the Carolingian Empire for decades. The monks had fled to Beauvais, taking
with them the casket in which the relics of St Vaast were preserved. Otto, at the
time bishop of the city, had received the community with honour and had ordered
that the relics be interred in the cathedral in order to be properly worshipped.
Now the danger seemed to be over, the site of the abbey had been repopulated
by monks coming from various parts of the region, the walls had been repaired,
and a new series of ditches had been excavated around them. However, the phy-
sical reoccupation of the site by the community had evidently not been enough
to bring the monastery back to its old glory. The absence of the saint from his
original resting place meant that the community was still destitute and the future
did not look hopeful. This was why the monks wanted to bring the body of their
patron back where they felt it belonged.
Understandably, when faced with the prospect of being deprived of such a
precious thing, the people of Beauvais turned a deaf ear. It took a lot of argument,
and the benign intervention of the new bishop of Beauvais, Honoratus, to con-
vince them to return the relics. The arrival to Arras of those revered remains must
11
‘Liber instrumentorum de possessionibus, rebus, sive dignitatibus, quas Casauriense
monasterium habuit, habet, vel habere debet’: Paris, Bibliothèque Nationale de France, fonds latin
5411, cc. 20 v –22 r; L. A. Muratori, Rerum Italicarum Scriptores, 25 vols (Milan, 1726), II.2, cols
767–1018 (col. 781).
Page 297
Ordernr. 061297 P1
282 Antonio Sennis
have taken place in an incredible scenario. According to the account of this

episode, most probably penned by a monk from St Vaast who eye-witnessed these
events, not only a multitude of clerics, but also the whole of the rural population
gathered to pray and worship at the homecoming of the sacred body.12
The place that a monastery occupied in the world was not simply physical, and
the reoccupation of the original site was not necessarily sufficient to bring a com-
munity back to its previous splendour. This is confirmed by the fact that the
material destruction of sacred places or objects, or the abandonment of the site,
could at times be invoked as the cause — real or claimed — of the loss of memo-
ries and traditions, in other words of the community’s identity. This is for exam-
ple the case of the monastery of Novalesa, in northern Italy (Fig. 13.1). According
to the chronicler who, in the first half of the eleventh century, collected the
traditions of his community, when the Saracens had attacked the abbey in the first
half of the tenth century the community had been obliged to flee. The site had
remained abandoned for some thirty years and this had resulted in such oblivion
that the monks could not even remember the location of the most venerated
graves. They therefore had to ask a woman, two hundred years old, to guide them
around the site. The very old woman is a narrative expedient symbolizing the fact
that the memory of individual places could indeed have been temporarily eclipsed
but that the link between the community and its past had not been completely cut
and it was possible to rediscover it, as it was possible to rediscover, sometimes to
invent, the sacred graves.13
The loss of memory, due to the abandonment of the original site, could indeed
have consequences more serious than those suffered by the monks of Novalesa. As
is well known, after the first destruction in 581 by the Lombards, the monks of
Montecassino stayed in Rome for more than a century (Fig. 13.1). Only in 717,
guided by Petronax, was monastic life re-established at Montecassino. Petronax,
helped by some hermits that had continued to live on the site, and by the three
12
Sermo de relatione corporis Beati Vedasti, in Monumenta Vedastina Minora, ed. O. Holder-
Egger, MGH SS, 15 (Hannover, 1887), pp.396–405 (pp.402–04, 404). Use of relics in Carolin-
gian Europe: P. Geary, Furta Sacra: Thefts of Relics in the Central Middle Ages, rev. edn with new
preface (Princeton, 1991). Relics as an attractive source of power: S. Boesch Gajano, ‘Reliques et
pouvoir’, in Les Reliques: Objects, Cultes, Symboles, ed. E. Bozóky and A. M. Helvétius (Turnhout,
1999), pp.255–69.
13
Cronaca di Novalesa, ed. G. C. Alessio (Turin, 1982), ch. 13, p.112. For the comparable case
of the monastery of Nonantola, see Translatio et miracula sanctorum Senesii et Theopontii, ed. P. E.
Schramm, MGH SS, 30.2 (Leipzig, 1934), pp.984–92 (p.987).
Page 298
Ordernr. 061297 P1
young Beneventans Paldo, Taso and Tato (who had just founded the abbey of
S. Vincenzo al Volturno), repaired the ruined buildings and gave a new start to the
community, whose identity was mainly founded on the fact that it was there, on
that precise site, that Saint Benedict was buried (even if, as I will discuss later,
nobody seemed to know exactly where). On the other hand, though, at least by
the beginning of the eighth century, a different tradition started to circulate.
According to this version of the events, while the community was in Rome, some
monks from the monastery of Fleury arrived at Montecassino to honour the burial
of the father of western monasticism. Once they arrived on site, the only thing
they found were ruins, and there was nobody who could tell them where the grave
was. Only by chance did these monks find the stone that identified the burial of
Benedict and of his sister Scholastica. They uncovered that abandoned tomb, took
the sacred bones, and brought them to France, to give them the burial they
deserved.14 According to the Fleury version, the transalpine monks had been right
in doing so because the monastery of Montecassino did not exist any more, since
its community had abandoned the site. And it is interesting to notice that those
hermits who, according to the Cassinese version of the events, had continued to
live on site represented the actual claim that the site had not been abandoned
completely, and that the physical space had continued to be an essential compo-
nent of the memorial identity of the community.
Spaces of Monks, Spaces of Power
A foundation, a prophetic vision, the happy ending of some relics’ exile: at first
sight the three episodes I have outlined might appear to have little in common and
to be hard to connect, set apart as they are by time and space. However, at a closer
look, at least the narratives in which these episodes are embedded do show some
consistency, some comparable features.15
The first thing that these three stories have in common is that, as I have
stressed, in one way or the other, and more or less explicitly, they all talk about
14
Translatio corporis s. Benedicti in Franciam, ed. R . Weber, Revue bénédictine, 62 (1952),
140–42. On the lack of knowledge of the location of Benedict’s tomb, see below, p.291.
15
On the intricate relationships between texts, the events they (purportedly) relate, and the
social contexts in which they come into being and circulate, from the point of view of a medie-
valist, see G. M. Spiegel, The Past as Text: The Theory and Practice of Medieval Historiography
(Baltimore, 1997).
Page 299
Ordernr. 061297 P1
284 Antonio Sennis
space. The space of which these three stories talk is, first of all, vividly physical: the
forests and fields of central Germany; the islands and waters of the extreme north;
a monastery and a city in north-eastern France. However, the space of which these
three stories talk is also, undeniably, a mental image: the power of a wilderness
that clearly stands for the eastern deserts of the monastic tradition; the blankness
of a landscape destitute of the Christian God; the desolation of an abbey filled
again with monks but still felt as empty, deprived as it is of the body of its patron.
Secondly, even if they relate events chronologically and geographically distant
one from the other, these three narratives were put together in a quite coherent
framework. The Life of Sturm — who, after founding Fulda, became its first
abbot — was written by one of his successors, Eigil, around 800; the Life of
Anskar was composed between 865 and 876; the story of the translation of the
relics of Saint Vaast was compiled in the immediate aftermath of the event, in the
early 890s.16
The period in which these narratives were composed can be defined as a second,
mature phase of the political, institutional and cultural parabola of the Carolingians.
After the conquests through which, in the second half of the eighth century, the
kingdom of the Franks had reached European dimensions, the expansion had at
first slowed down and then stopped; the territory had been reorganized; new
missionary waves had endeavoured to shift the frontiers of Christianity to the
north and to the east. The political, economic, cultural and religious horizon had
thus adjusted to include other counterparts, other allies, other enemies.17
It is difficult to say whether the establishment of this new, wider scenario had
a direct influence on how people, or at least the members of the intellectual elites,
looked at the world around them. What can be said, though, is that it is precisely
in those decades that a transformation of the ways in which space was represented
seems to have occurred.
16
See, respectively, Engelbert, Die Vita Sturmi, pp.16–20; I. Wood, The Missionary Life:
Saints and the Evangelisation of Europe 400–1050 (London, 2001), pp.52, 125; Sermo de relatione
corporis Beati Vedasti, p.396.
17
On the period see, in general, P. Godman and R . Collins, eds, Charlemagne’s Heir: New
Perspectives on the Reign of Louis the Pious (814–840) (Oxford, 1990); J. Nelson, Charles the Bald
(London, 1992); S. MacLean, Kingship and Politics in the Late Ninth Century: Charles the Fat
and the End of the Carolingian Empire (Cambridge, 2003). Frankish missions of the eighth and
ninth centuries: R . Fletcher, The Conversion of Europe: From Paganism to Christianity 371–1386
AD (London, 1997), pp.193–284; and, with a particular focus on the narratives of the missions,
Wood, Missionary Life, pp.57–141.
Page 300
Ordernr. 061297 P1
For example when, between 842 and 847, Hraban Maur devoted to geography
books XI–XIII of his De rerum naturis (a work later known as De universo), he
drew on Isidore’s Etymologiae but consistently interpreted the latter’s descriptions
using the transcendent and mystical categories of the sensus spiritualis.18 On the
other hand, the works of cosmographers such as Dicuil (whose De mensura orbis
terrae was written in 825 and dedicated to Louis the Pious) and the anonymous
monk author of the De situ orbis (active at Ferriéres or Auxerre between 856 and
870) demonstrate that the geographical picture that was presented was not the
one inherited from classical antiquity. Rather, what was described was a new and
reorganized world, one in which the role of the Mediterranean as structural axis
was secondary to that assigned to the Ocean. These monks, whose works were
conceived to be employed in the teaching of geography, still used ancient sources
but altered the information they found in them in order to include the islands of
northern Europe, ignore the lands south and east of the Mediterranean, and
strongly emphasize the navigability of the Ocean. Moreover, both authors proved
to be quite aware of what can be defined as the geo-politics of the late Carolingian
world and in more than one point of their works stressed the impact of the
Vikings on the, human as well as physical, geography of northern Europe. The
author of the De situ orbis even explicitly stated in his prologue that he had been
prompted to get down to work by some members of his community who wanted
to have a clearer idea of the routes followed by the Vikings for their incursions.19
Lastly, all these three narratives are the outcome of a monastic culture. Even
if, strictly speaking, it was produced in the episcopal milieu of Hamburg-Bremen,
18
Rabani Mauri. De universo libri viginti duo, ed. J. P. Migne, Patrologia Latina, 111 (Paris,
1841–64), cols 9–614 (cols 309–76). Hraban as reader and utilizer of Isidore: J. Fontaine, ‘Isidore
de Séville et la mutation de l’encyclopédisme antique’, in Tradition et actualité chez Isidore de
Séville, ed. J. Fontaine, special issue, Cahiers d’histoire mondiale / Journal of World History /
Cuadernos del Historia Mundial, 9.3 (1966), 519–38 (pp. 536–37). Change of title from De
rerum naturis to De universo: E. Heyse, Hrabanus Maurus Enzyklopädie ‘De rerum naturis’
(Munich, 1969), pp. 2–3.
19
De mensura orbis terrae, ed. J. Tierney and L. Bieler, Scriptores Latini Hiberniae, 6 (Dublin,
1967) pp.44–103; De situ orbis, ed. R . Quadri (Padua, 1974). On these issues, see P. Gaultier
Dalché, ‘Tradition et renouvellement dans la représentation de l’espace géographique au IX e siècle’,
Studi medievali, 3rd series, 24 (1983), 121–65 (pp.142–65). A different view is taken by
N. Lozovsky, The Earth is Our Book: Geographical Knowledge in the Latin West ca. 400–1000
(Ann Arbor, 2000), pp.154–55, where it is argued that the geographical knowledge of this period
essentially reproduced the image of the world that had been familiar since antiquity, and tended
to ignore contemporary reality.
Page 301
Ordernr. 061297 P1
286 Antonio Sennis
the Vita Anskarii can in fact also be considered as the product of a fully monastic
context. Before becoming bishop of Bremen and then archbishop of Hamburg,
Anskar had completed all his education and career first at Corbie and then at
Corvey. Rimbert, his biographer as well as successor on the Hamburg see, had also
been educated in a monastery, Turholt. Moreover, while undertaking his narrative
task he was in close contact with the community at Corbie, to which he dedicated
his work, and with one member in particular, the predestinatarian theologian
Ratramnus.
In the evolution of the social, economic, political, cultural and religious
structures that occurred in early medieval Europe, those monastic communities
were sometimes the protagonists, sometimes the victims, most of the time the
narrators, always the guardians of memory.20 Besides, those places of worship were
always in close relation to different forms of powers, and their members were
among the protagonists of change — in political, institutional and social struc-
tures, in models of territorial identity, in the organization of landowning, for
example. The focal role that monasteries had with respect to the surrounding
society is mirrored in the fact that monastic sources allow us to relate the changes
in land organization to changes in the mental structures of society, and in the way
people perceived the space around them.21
20
Absolute pre-eminence of churches, and especially monasteries, in the transmission of texts
and documents in early medieval Europe: P. Cammarosano, Italia medievale: Geografia e storia
delle fonti scritte (Rome, 1991), pp.39–111. Anskar’s strong monastic profile: W. Haas, ‘Foris
apostolus – intus monachus: Ansgar als Mönch und “Apostel des Nordens”’, Journal of Medieval
History, 11 (1985), 1–30. Rimbert’s upbringing and career: Vita Rimberti, ed. G. Waitz, MGH
SRG in usum scholarum, 55 (Hannover, 1884), pp.80–100.
21
A. Guerreau, ‘Le champ sémantique de l’espace dans la vita de saint Maieul (Cluny, début
du XIe siècle)’, Journal des Savants (1997), 363–420. C. Sonnlechner, ‘The Establishment of New
Units of Production in Carolingian Times: Making Early Medieval Sources Relevant for Environ-
mental History’, Viator, 35 (2004), 21–48. Centrality of a monastery within a given society: B. H.
Rosenwein, To Be the Neighbor of Saint Peter: The Social Meaning of Cluny’s Property 909–1049
(Ithaca and London, 1989). In this essay I use evidence from male communities, mainly owing
to the nature of the sources I am more familiar with. On the relations between gender and exterior
space, see S. S. Morrison, Women Pilgrims in Late Medieval England: Private Piety as Public
Performance (London, 2000), pp.83–93; K. Heene, ‘Gender and Mobility in the Low Countries:
Travelling Women in Thirteenth-Century Exempla and Saints Lives’, in The Texture of Society:
Medieval Women in the Southern Low Countries, ed. E. E. Kittell and M. A. Suydam (London,
2004), pp.31–49. R . Gilchrist, Gender and Material Culture: The Archaeology of Religious Women
(London and New York, 1994), discusses monastic concepts and symbols of (mainly interior)
space with regard to female communities.
Page 302
Ordernr. 061297 P1
Space as Discourse
It has to be said that the use of space as a central element of a system of representa-
tion is not, in principle, exclusive either to early medieval monastic communities
or, more generally, to religious elites.22 The need to interpret geography in the
framework of dynamics of culture, and power, and the use of texts as media to
convey an explanation of the landscape as part of a political strategy are in fact
quite pervasive phenomena, common to different social groups and to different
historical contexts. Space is indeed a cultural image.23 To take just one example,
from the seventh century onwards the creation of royal forests concretely pro-
jected on the physical landscape the polarization laics and clerics created by the
rituals of hunting, from which the latter had been barred since the beginning of
the sixth century. This projection, codified in writing by the law, was no less
mental than material: the areas reserved to royal hunting were in fact not always
necessarily fenced, and the demarcation of the space often relied on social
dynamics.24
This is also true if we consider the dynamics through which single elements,
or portions, of the space become sacred. Most societies, in order to define their
own distinctive sacred topography, tend in fact to look at different parts of the
surrounding space, not only monuments, and draw upon memory and tradition.25
22
The most seminal examples of how, in different contexts, the articulation of space can be
interpreted, on the one hand, as a symbolic ordering of the world and, on the other, as an
expression of institutional interests are, respectively, P. Bourdieu, ‘The Kabyle House or the
World Reversed’, in Bourdieu, Algeria 1960 (Cambridge, 1979), pp.133–53; and M. Foucault,
Discipline and Punish: The Birth of the Prison (Harmondsworth, 1977). See also M. Foucault,
‘Espace, savoir et pouvoir’, in Foucault, Dits et Écrit 1954–1988, 4 vols (Paris, 1994), IV , 270–85.
23
D. Cosgrove and S. Daniels, eds, The Iconography of Landscape (Cambridge, 1988), p.1. See
also J. B. Harley, ‘Maps, Knowledge, and Power’, in ibid., pp.277–312.
24
Wickham, ‘European Forests in the Early Middle Ages’, pp.160–61, cites the relevant laws.
See also R . Hennebicque, ‘Espaces sauvages et chasses royales dans le Nord de la France, VIIe– IX e
siècle’, Revue du Nord, 62 (1980), 35–60. The bipartition of hunting (hunting with dogs/hunting
with falcons) has been recently put in symbolic relation with the bipartition of the scenario in
which it took place (inside/outside the forest): hunting would have contributed to the polariza-
tion of space in the Middle Ages. See A. Guerreau, ‘Chasse’, in Dictionnaire Raisonné de l’Occident
Médiéval, ed. J. Le Goff and J.-C. Schmitt (Paris, 1999), pp.166–78, esp. pp.168–72.
25
D. O’Sullivan, ‘Space, Silence and Shortage on Lindisfarne: The Archaeology of Ascetism’,
in Image and Power in the Archaeology of Early Medieval Britain: Essays in Honour of Rosemary
Cramp, ed. H. Hamerow and A. MacGregor (Oxford, 2001), pp.33–52 (p.36).
Page 303
Ordernr. 061297 P1
288 Antonio Sennis
However, the perception of space and its general system of representation did
change in the transition from antiquity to the Middle Ages, largely due to the in-
fluence of the developing Christian system of thought. For medieval clerics the world
was in fact not only the physical reality but was also constituted by the whole of
human history: space and time were inextricably tied.26 Therefore, for example,
in the seventh century the Irish monk Adomnán, in relating a prophecy regarding
Columba’s fame, explicitly linked the end of time to the end of the known land.
Moreover, in the early Middle Ages particularly large world maps aimed at portraying
the course of universal history in symbiosis with the whole of historical space.
This mental framework, in which time and space are indissolubly bound
together, is further revealed by the widespread circulation, in the Middle Ages, of
the legend of Alexander the Great. In this romance Alexander imprisoned Gog
and Magog until the end of time on what was then conceived as the ultimate
inviolable barrier: the Caucasus. For the enclosed nations geographical and
temporal prisons therefore coincided.27
This progressive permeation of space and time was one of the crucial develop-
ments of Western Christian mentality, a system of thought that, for at least the
first three centuries of its history, was strongly hostile to the idea that any physical
site should be venerated. It was in the course of the fourth century that the
concept of a Christian sacred topography slowly began to catch on. The tradi-
tional importance of the past in the framework of the Christian faith was thus
combined with the developing need to identify, physically as well as ethically, the
Church of the present with the Church of the early martyrs. This resulted in the
sacralization of specific places, that is the materialization, in some definite points
of the landscape, of the sacredness of time.28 It was indeed western monasticism,
26
A. Guerreau, L’avenir d’un passé incertain: Quelle histoire du Moyen Âge au XXIe siècle? (Paris,
2001), p.233.
27
Adomnán: T. O’Loughlin, ‘Living in the Ocean’, in Studies in the Cult of Columba, ed.
C. Bourke (Dublin, 1997), pp.11–23 (p.20). Medieval maps: A. D. von den Brincken, ‘Mappa
Mundi und Chronographia’, Deutsches Archiv für Erforschung des Mittelalters, 24 (1968), 118–86;
and E. Edson, Mapping Time and Space: How Medieval Mapmakers Viewed their World (London,
1997). Legend of Alexander: A. H. Anderson, Alexander’s Gate: Gog and Magog and the Inclosed
Nations (Cambridge, MA, 1932). Caucasus as an inviolable limit: A. Giardina, ‘Roma e il
Caucaso’, Settimane di studio del centro italiano di studi sull’alto medioevo (Il Caucaso: cerniera fra
culture dal Mediterraneo alla Persia (secoli IV – XI)), 43 (1996), 85–141 (pp.86–89).
28
R . A. Markus, ‘How on Earth Could Places Become Holy? Origins of the Christian Idea
of Holy Places’, Journal of Early Christian Studies, 2 (1994), 257–71 (pp.268–71), arguing against
Mircea Eliade’s idea that it is almost tautological that every religion provides itself with sacred
Page 304
Ordernr. 061297 P1
especially in the context of its prescriptive texts, which really redefined the notion
of space.29
The various monastic rules provide in fact evidence of the development of a
new sense of space and of a quite original effort to modify the physical framework
of social life. Moreover, in the transition from the texts concerning the earlier
desert fathers to the cenobitic rules there seems to be a clear development, a
progressive process of construction of represented space. This development did
not occur without contradictions, and in the Carolingian and post-Carolingian
epochs, monastic chroniclers seem, at times, to have been aware of the fact that
too long and detailed a description of the landscape surrounding their community
could indeed contravene the biblical tradition, by putting too much emphasis on
the place in relation to the community.30 However, in the course of time the
physical desert of the eastern fathers became an ideological scenario for western
monks, and it is indeed significant that what can be defined as the ‘myth of the
desert’ enjoyed a long-lasting fortune in the Middle Ages becoming, for example,
a cornerstone of Cistercian ideology.31
Space and Memory
At times, a community’s capability to control the processes of description, inter-

pretation and memorization of space could be more avowed than actual. Among
the wave of immigrants who, between the 840s and 870s, followed the imperial
sites. For a critique of Eliade’s approach, and for an analysis of the fundamental dichotomy
between locative and utopian (and, as such, non-locative) orientations in religion, see also J. Z.
Smith, Map Is Not Territory (Leiden, 1978), esp. pp.1–23.
29
J. Le Goff, ‘Discorso di chiusura’, Settimane di studio del centro italiano di studi sull’alto
medioevo (Popoli e paesi nella cultura altomedievale), 29 (1983), 805–38 (p.823).
30
See, for example, Ex miraculis et translationibus s. Bavonis, ed. Holder-Egger, MGH SS,
15.2, pp.589–99 (p.591). (The author quotes II Maccabees 5. 19: ‘But the Lord did not choose
the nation for the sake of the holy place, but the place for the sake of the nation.’)
31
B. Ward, ‘The Desert Myth: Reflections on the Desert Ideal in Early Cistercian Monasti-
cism’, in One Yet Two: Monastic Tradition in East and West, Orthodox-Cistercian Symposium
(Kalamazoo, 1976), pp.183–99. Ideological use of the eastern deserts in western early monasti-
cism: H. Löwe, ‘Westliche Peregrination und Mission: Ihr Zusammenhang mit den Länder und
Völkerurkundlichen Kenntnissen des früheren Mittelalters’, in Settimane, 29 (1983), 327–72;
P. Noisette, ‘Usages et représentations de l’espace dans la Regula Benedicti: Une nouvelle
approche des significations historiques de la Règle’, Regulae Benedicti Studia: Annuarium
internationale, 14–15 (1985–86), 69–80 (pp.69–70).
Page 305
Ordernr. 061297 P1
290 Antonio Sennis
expeditions in central Italy, Frankish and Alamannic settlers continued, for

example, to describe the land they were acquiring according to an administrative
vocabulary that they had learnt in their places of origin, regardless of the style used
by the monastery they were dealing with, S. Clemente at Casauria.32 And monas-
teries could also be obliged to turn to local communities to make up for their lack
of knowledge of the territory they controlled. Therefore when, between the end
of the twelfth and the beginning of the thirteenth centuries, Rainaldus, abbot of
the monastery of S. Salvatore alla Maiella, some 150 km east of Rome, decided to
commission the compilation of a cartulary, the community knew practically
nothing about the consistency and articulation of its revenues (Fig. 13.1). The in-
formation necessary in order to accomplish the cartulary had to come from sworn
testimonies of the inhabitants of the communities surrounding the monastery.33
This lack of knowledge could even extend to the internal spaces of a monas-
tery, and in some cases the community seems not to have any idea about the
location of illustrious graves, which one would expect to be the memorial focus
of a monastery, at least in principle. For example, at the end of the tenth century
the author of the chronicle of the monastery of Saint Andrew on Monte Soratte,
some kilometres north of Rome, wrote of how his community had once been
visited by Charlemagne, shortly after his imperial coronation (Fig. 13.1). Accord-
ing to the monk, the emperor donated to the community some relics of Saint
Andrew, and these relics were put in a grave and venerated. However, he could
not tell where this grave was.34
32
Examples (all dated between 873 and 880): Allo ‘ex genere Francorum’ who declares
himself ‘habitator in pago Teatense’ (‘Liber instrumentorum’, cols 86 r–86 v); Sisenandus ‘ex genere
Francorum’ who declares that the properties he is selling are located ‘in loco qui dicitur Casauria,
in pago Pinnensi’ (ibid., cols 75 v –76 r); Ainardus ‘ex genere Alamannorum’ who declares himself
‘habitator in pago Teatense’ (ibid., col. 91 r); Johannes ‘ex genere Alamannorum’ who declares
himself ‘habitator in pago Pinnense’ (ibid., col. 93 r); Rimo ‘ex genere Alamannorum’, who declares
himself ‘habitator in pago Pinnense’ (ibid., col. 97 r, and cols 99 r–99 v); Sanson ‘ex genere Franco-
rum’ who declares himself ‘habitator in pago Balvense’ (ibid., cols 114 r–114 v); Ildeprand ‘ex genere
Francorum’ who declares himself ‘habitator in pago teatino’ (ibid., cols 115 v –116 r). Pagus is not
a term commonly used to describe districts in central Italy. On transalpine immigrants to Italy in
the early Middle Ages, see E. Hlawitschka, Franken, Alemannen, Bayern und Burgunder in
Oberitalien (774–962): Zum Verständnis der fränkischen Königsherrschaft in Italien (Freiburg im
Breisgau, 1960).
33
Biblioteca Apostolica Vaticana, Archivio di S. Pietro, Caps. 72, fasc. 53, col. 29 (partial
edition of the manuscript in Collectionis Bullarum Sacrosantae Basilicae Vaticanae (Rome 1747),
pp.i–liii).
34
Il Chronicon di Benedetto di S. Andrea del Soratte e il Libellus de imperatoria potestate in
urbe Roma, ed. G. Zucchetti (Rome, 1920), p.116.
Page 306
Ordernr. 061297 P1
Sometimes this lack of knowledge is even more striking. At Montecassino, the

place where Saint Benedict was allegedly buried was accidentally found during the
great renovation of the abbey church started, in 1066, by Abbot Desiderius.
Everything suggests that the discovery was casual, and in fact Desiderius had to
modify the original plan of the basilica that, in the final version, had an elevated
presbytery in order not to touch the sacred remains.35 On the other hand, when
Emperor Henry II visited the monastery in June 1022 and expressed doubts on
the fact that Saint Benedict was really buried at Montecassino and not, as claimed
by a quite credible tradition, in the French monastery of Fleury, the community
could not say anything and the emperor was convinced only when, during a night
of torment due to what appears to have been colic, he had a vision of the saint who
told him that his pain would stop only if he was convinced that the monks of
Fleury were lying about his burial.36
The monks of the other great Italian imperial abbey, S. Clemente at Casauria,
experienced similar problems. There the community claimed that, in founding the
monastery in the 870s, Emperor Louis II had buried the relics of Saint Clement
somewhere, but they did not know where. It took them a long time, and lots of
miraculous appearances by the saint himself, to convince a series of papal legates
who wanted, in the 1150s, to establish if the saint was buried there or in the
basilica of Saint Clement in Rome.37
Certainly, these types of oblivion could be strategically concocted in order to
sustain otherwise groundless claims: the monks of Casauria had to deal with a —
for them — very unfavourable Translatio Sancti Clementis written by no less a
person than Leo of Ostia, who claimed that the relics of the martyr had never left
Rome; maybe, when Henry II arrived at Montecassino, the body of Saint Benedict
had indeed been long moved to Fleury; maybe the monastery on the Monte
Soratte had never possessed any relics of Saint Andrew. What is important to
stress is the fact that the coexistence of a more or less justified liturgy of memory
35
Chronica Monasterii Casinensis, III.26, ed. H. Hoffmann, MGH SS, 24 (Hannover, 1980),
p.395. See also Il sepolcro di S. Benedetto, Miscellanea cassinese, 27 (Montecassino, 1951);
P. Meyvaert, ‘Peter the Deacon and the Tomb of Saint Benedict’, Revue bénédictine, 65 (1955),
3–70.
36
Chronica Monasterii Casinensis, II.43, ed. Hoffmann, pp.247–49.
37
A. Pratesi, ‘Ubi corpus beati Clementis papae et martyris requiescit’, in Contributi per una
storia dell’Abruzzo adriatico nel medioevo (Chieti, 1992), pp.115–31.
Page 307
Ordernr. 061297 P1
292 Antonio Sennis
and a more or less opportunistic caducity of memories appears to be normal and

dealt with with self-confident awareness.38
Complex Spaces
The interaction between community and space could take place at the highest
possible level, such as the one constituted by the intellectual crème of the
Carolingian elites. In that case, the result could be a real turning point in western
civilization. The space that the Carolingian monks of the ninth century observed,
remembered and imagined was a new world. It was new because the political
scenario had changed, but it was new also because it was described and understood
in new ways, using new cultural categories.
Even when the interaction occurred at a more local level, monastic communi-
ties not only used the space around them, but also placed their interpretations of
it in the context of a rich and articulated set of practices of liturgical memory from
which their members could, and did, draw systematically and collectively, in order
to define the present and carve out for themselves spheres of dominion. Those
shared interpretations of the surrounding space, and the apparatus of practices
which substantiated and perpetuated them, could be arranged in a real theology
of memory and oblivion, through the liturgical prayer of intercession for the living
and the dead.39
Space pre-exists, and is independent from, human experience. However, it is
qualified by people, explained by their language, organized by their activity. In
other words, by determining spatial categories (or, to put it more simply, by decid-
ing what the ‘proper place’ for things is), most societies also create and qualify cul-
tural concepts. These concepts have a strong linguistic dimension but, in general,
spatial perceptions rely on a system of representations of space that is more articu-
lated than the one coded in language.40
38
Emphasis on the imaginative component of this set of perceptions, particularly with regard
to the community’s past, is placed by Remensnyder, Remembering Kings Past, pp.1–4.
39
Relations between community, commemoration and memory: O. G. Oexle, ‘Memoria und
Memorialüberlieferung im früheren Mittelalter’, in Frühmittelalterliche Studien, 10 (1976),
70–95. See also P. Geary, Phantoms of Remembrance: Memory and Oblivion at the End of the First
Millennium (Princeton, 1995); Remensnyder, Remembering Kings Past.
40
M. Douglas, Purity and Danger: An Analysis of the Concepts of Pollution and Taboo, new
edn (London, 2002), esp. pp.44–50; Cantarella, ‘Lo spazio dei monaci’, p.808. Morrison, Women
Page 308
Ordernr. 061297 P1
Today the term space corresponds, in its current use, to the notion of
Euclidean space. This is not the case for the Middle Ages, because there is no
medieval word that can be considered its equivalent: spatium meant interval, dis-
tance, nothing more.41 On the other hand, as we have seen, a complex and articu-
lated perception of space not only existed in the Middle Ages, but was entirely
central to people’s mentality. At Iona, for example, the sea was a fundamental
element of the community’s spiritual and ideological sense of identity, and at
Lindisfarne the whole surrounding landscape was interpreted according to the
principles of biblical exegesis.
In the mind of late antique and early medieval clerics every element of the
world that senses could experience was, in other words, also a symbol of an
invisible reality.42 Moreover, space (or, better, the control of it, in all senses) was
then, as it is now, indispensable to any exercise of power.
Pilgrims in Late Medieval England, p.84. Language and space: S. C. Lavinson, ‘Space: Linguistic
Expression’, in International Encyclopedia of the Social and Behavioral Sciences, 29 vols (Amster-
dam, 2001), XXII, 14749–52 (p.14749). See also L. Talmy, ‘How Language Structures Space’, in
Spatial Orientation: Theory, Research, and Application, ed. H. L. Pick, Jr, and L. P. Acredolo (New
York, 1983), pp.225–82; S. Svorou, The Grammar of Space (Amsterdam, 1994), pp.1–40.
41
A. Guerreau, ‘Structure et évolution des représentations de l’espace dans le haut Moyen Âge
occidental’, Settimane, 50 (2003), 91–115 (pp.93–94). At times the term spatium could be
explicitly referred to temporal determinations, as it also happens nowadays. For some Italian
examples, see Chartularium Cupersanense, ed. D. Morea (Montecassino, 1892), I, no. 2 (May
889), pp.3–5 (p.5); Codice diplomatico del monastero benedettino di S. Maria di Tremiti
(1005–1237), ed. A. Petrucci (Rome, 1960), II, no. 17 (October 1035), pp.57–59 (p.59);
Chronica Monasterii Casinensis, ed. Hoffmann, pp.129, 327, 522, 557. On the relations between
space and time, see above, p.288. Use of spatial references to express units of time in the late
Middle Ages: J. Paul, ‘Expression et perception du temps d’après l’enquête sur les miracles de Louis
d’Anjou’, in Temps, mémoire, tradition au Moyen Âge (Aix-en-Provence, 1983), pp.19–41
(pp.33–34); and R . Bartlett, The Hanged Man: A Story of Miracle, Memory and Colonialism in
the Middle Ages (Princeton, 2004), pp.63–67.
42
G. de Nie, ‘Seeing and Believing in the Early Middle Ages: A Preliminary Investigation’, in
de Nie, Word, Image and Experience: Dynamics of Miracle and Self-Perception in Sixth-Century
Gaul (Aldershot, 2003), ch. VII, p.69. For Iona and Lindisfarne, see, respectively, O’Loughlin,
‘Living in the Ocean’; O’Sullivan, ‘Space, Silence and Shortage on Lindisfarne’. On monastic
space, see also, for a later period, M. Cassidy-Welch, Monastic Spaces and their Meanings:
Thirteenth-Century English Cistercian Monasteries (Turnhout, 2001). Some interesting ideas may
be found in D. Harrison, Medieval Space: The Extent of Microspatial Knowledge in Western Europe
during the Middle Ages (Lund, 1996). For a more general discussion, see also G.-G. Granger, La
pensée de l’espace (Paris, 1999) and the essays collected in F. Remotti, P. Scarduelli and U. Fabietti,
eds, Centri, ritualità, potere: Significati antropologici dello spazio (Bologna, 1989).
Page 309
Ordernr. 061297 P1
294 Antonio Sennis
It is this complexity of values that helps explain how, and why, those monastic
communities underpinned their efforts of social supremacy by constantly
drawing, rightly or wrongly, on their vocation to turn space into memory,
memory into texts, texts into power.43
43
For space as a fundamental component of power, see M. Foucault, ‘The Eye of Power’, in
Foucault, Power/Knowledge: Selected Interviews and Other Writings 1972–1977 (New York,
1980), pp.146–65 (p.149). I am grateful to Wendy Davies, Paul Fouracre, David d’Avray, Nick
Everett and Chris Wickham for their comments on an earlier version of this paper. They are not
responsible for any of my mistakes or shortcomings.
Page 310
Ordernr. 061297 P1
P OPULATIONS, T ERRITORY AND C OMMUNITY

M EMBERSHIP: C ONTRASTS AND C ONCLUSIONS
Wendy Davies
S
haring approaches to people and space in different parts of western Europe
in the Middle Ages has been a stimulating and enriching experience for
those who have worked together on this book; and it has had some un-
expected results. One of the most striking of the latter is the influence of national
historiographical traditions. The Spanish framework is dominated by progress
towards the feudalization of society and the emergence of a well-grounded
military aristocracy in the tenth and eleventh centuries; it is also immensely influ-
enced by the backlash against the eighth-century depopulation theories which
characterized much twentieth-century Spanish writing. Hence, those who now
write tend to emphasize long continuities and the social antecedents of those
emerging aristocracies; and there is, on the whole, more interest in hierarchy than
in community. The English framework has its search for continuities too: it is
especially marked by the belief that pre-existing large units of ownership split to
form smaller estates, notably in the ninth and tenth centuries (called ‘multiple
estates’ by some writers, although not by those contributing here 1), and by the
belief that parish* boundaries reflect estate boundaries. Those who write now,
therefore, hesitate to investigate the relationship between church community and
proprietor’s community because there is a strong belief that they were cotermin-
ous. The Icelandic writers in this book are more noticeable for denying the
1
For example, R . Fleming, ‘Lords and Labour’, in From the Vikings to the Normans: The Short
Oxford History of the British Isles (800–1100), ed. W. Davies (Oxford, 2003), pp.107–37
(pp.108–11); Glanville Jones was the most forceful exponent of this theory, for example in G. R . J.
Jones, ‘Multiple Estates and Early Settlement’, in Medieval Settlement: Continuity and Change,
ed. P. H. Sawyer (London, 1976), pp.15–40.
Page 311
Ordernr. 061297 P1
296 Wendy Davies
predominating Icelandic historiography than perpetuating it; these contributors

explicitly strive to get away from traditional approaches, opposing especially the
concept that the settlement of Iceland occurred through the initiatives of free
farmers, exercising something like democratic rights. They therefore tend to em-
phasize the initiatives of chiefs, and devote more time than most other contribut-
ors to methodology. The French and Frankish pieces are an exception to this
trend and fall less obviously within any national historiographical framework —
perhaps because the writers are English. Given the strength and distinctiveness of
most of these national approaches, we need to be alert to the distinction between
genuine regional differences and those that are simply the product of different
national perceptions.
All contributors have a concern with territory and they deal with it in different
ways. Two essays in particular are concerned with physical markers, those of
Reynolds and Langlands on the English Wansdyke and of Lárusdóttir on earth-
works in north-east Iceland. The former argue strongly for the interpretation of
Wansdyke as a political frontier of the West Saxon kingdom, the massive linear
earthwork being an expression of political capacity to control labour. The latter
argues for earthworks in Reykjahverfi as separators between (early) farms, as also
between infield and outfield within farms; some of the earthworks went out of use
at a relatively early stage as some farmsteads were abandoned and property was
consolidated, perhaps as a result of planned abandonment. It would not be
appropriate to propose either regional similarity or regional difference on the basis
of these two papers, but there is enough to note both political and proprietary uses
of earthworks in our period. That fact is more important than it might seem:
linear features in the landscape can arise for both kinds of reason, marking
political frontiers and property boundaries; indeed, to the extent that early
Icelandic earthworks marked chieftains’ property, they might mark both together;
but they do not necessarily have to do with any kind of local community. Others,
Vésteinsson and Escalona especially, have emphasized that earthworks and linear
features are in fact unusual: many property boundaries utilized natural features,
whether to join or to separate, although turf or stone ‘cairns’ marked corners in
Iceland, roads are well-known as boundary markers in Spanish Castile, and crosses
might be used almost anywhere. The natural features included the streams,
springs, ditches, trees, woods, stones, ridges, meadows and heaths that are so
familiar from the boundary clauses of Anglo-Saxon and Welsh charters, but are
not so consistently recorded elsewhere. Indeed, for those who made records of the
property they owned, having a sense of physical space was extremely important,
Page 312
Ordernr. 061297 P1
POPULATIONS, TERRITORY AND COMMUNITY MEMBERSHIP 297
Photo 14.1. Arnarklettur (‘Eagle’s rock’), a natural boundary marker between the
farm Hreðavatn in the commune of Norðurárdalur and the farm Laxfoss in the
commune of Stafholtstungur, in Borgarfjörður, West Iceland. The stone is a part
of a lava field that forms the natural boundary between the communes and
properties. Photo: Andrew Reynolds.
even if in practice it was not particularly well informed.2 Markers of community

space, however, except through the use of the markers of constituent property
units (or in the very specialized case of those of a religious community’s ‘protected
space’3), are much less obvious and do not emerge strongly from these studies.4
2
See above, pp.279–81.
3
See W. Davies, ‘“Protected Space” in Britain and Ireland in the Middle Ages’, in Scotland
in Dark Age Britain, ed. B. E. Crawford (St Andrews, 1996), pp.1–19 (pp.4–7).
4
Bassett does however argue that fixed visible boundaries identified the territory of each
community; above, p.116.
Page 313
Ordernr. 061297 P1
298 Wendy Davies
There are few clear parallels to the distinctive land-use at the margins of Breton
plebes and communes.5
Very few communities in the early Middle Ages had anything like the legal
definition of communities of the Roman Empire, as Adela Cepas so clearly
demonstrates. Indeed, few early medieval communities beyond the lay or ecclesi-
astical household leave any evidence that their members had any sense of group
participation or that the group performed this or that function for its members;
in this the ninth-century Breton village communities discussed in the Introduct-
ion seem to be unusual, although we know of comparable communities from the
Middle Rhineland.6 The very fact that the strong and formal identities of the
Roman period disintegrated is itself significant.7 Indeed, Chris Wickham has
recently drawn attention to the difference between eastern and western rural
communities in this respect in the fifth to eighth centuries: while the villages of
Anatolia and Egypt have left plentiful evidence of their vitality as functioning
communities, with collective action and village officials, many of the settlements
of western Europe leave a weak sense of community and a stronger sense of
patronage networks at that time.8 Of course, some texts suggest that individuals
could feel a sense of belonging to groups beyond the household — to ‘peoples’ like
Visigoths or Franks or to smaller population groups like the English Hwicce or
Irish Osraige — but whether or not that sense was shared by all members of the
group is highly debatable, despite one suggestion that all the Franks got together
in the eighth century;9 it is an open question whether such groups had any prac-
tical social functions or any part in the regulation of social life. As for any shared
sense of belonging: it has recently been convincingly demonstrated that it is
5
6
See M. Innes, State and Society in the Early Middle Ages: The Middle Rhine Valley,
400–1000 (Cambridge, 2000), especially pp.94–140.
7
Cf. W. Pohl, ‘The Construction of Communities and the Persistence of Paradox: An
Introduction’, in The Construction of Communities in the Early Middle Ages: Texts, Resources and
Artefacts, ed. R . Corradini, M. Diesenberger and H. Reimitz (Leiden, 2003), pp.1–15 (p.12), on
the decline of civic identity in the post-Roman period in Europe. See also above, pp.41–42,
205–07, 214.
8
C. J. Wickham, Framing the Early Middle Ages (Oxford, 2005), ch. 7, especially pp.390–93,
397–98, 438–40, also ch. 8, pp.514, 516–18. Iñaki Martín Viso points out to me that the more
active presence of the state in the East required local communities to enter into dialogue and
thereby the vitality of their functions is more easily demonstrable.
9
Pohl, ‘Construction of Communities’, p.10.
Page 314
Ordernr. 061297 P1
difficult to attach any sense of belonging to anyone other than a person of high
status as also that the meaning of names like ‘Frank’ changed over time; ‘identities
had to be flexible and largely virtual’; ethnicities might be created, dismantled and
recreated.10 Functioning communities beyond the residential household did begin
to become clear in the West as they established distinctive identities in the central
Middle Ages, like the self-conscious identities of urban and craft associations.
Grenville Astill shows how late it was before urban identities formed in southern
England, although tithing groups should at least in theory have been clear from
the tenth century and guilds are certainly well evidenced then — he cites the long
distances that guild members might travel for guild purposes, up to 25 km and
exceptionally 80 km.11 We can therefore identify a series of important, related
problems: should we consider the possibility that, until the formation of distinc-
tive new communities like tithings or guilds, there were no communities beyond
the household or family? Were the Rhineland communities and the Breton plebes,
in essence village communities, truly unusual? Or should we suppose that rural
society changed in the West in the central Middle Ages and that the Rhenish and
Breton societies are precocious symptoms of a wider change, symptoms that
happen to have been well recorded?12 Alternatively, should we suppose that local
communities existed but are concealed in the record — communities of economic
cooperation, communities of quasi-legal competence, communities of believers,
and so on, at the same time overlaid by the networks of chosen associations (of
clientship, of tenancy, of exchange, of guild)? In practice most contributors to this
book adopt the latter assumption. The problem then becomes one of reconstruc-
tion: how can we spot the communities that must have been there? What criteria
should we use? How can we avoid circularity?
10
R. Wenskus, Stammesbildung und Verfassung: Das Werden der frühmittelalterlichen Gentes
(Cologne, 1961) for the classic statement; W. Pohl, ‘Telling the Difference: Signs of Ethnic
Identity’, and I. Wood, ‘Conclusion: Strategies of Distinction’, both in Strategies of Distinction:
The Construction of Ethnic Communities, 300–800, ed. W. Pohl and H. Reimitz (Leiden, 1998),
pp.17–69 and 297–303 (pp.21–22, 299, 63), for recent work. Cf. P. J. Geary, The Myth of
Nations: the Medieval Origins of Europe (Princeton, 2002), for the ‘myth’ of ethnogenesis,
especially pp.157–74.
11
See above, p.252.
12
Cf. Wickham, Framing, who argues forcefully that the weak communities in the pre-ninth-
century West developed much stronger collective identities and capacity for action in subsequent
centuries; cf. Fouracre, above pp.272–73. Halsall would see a contrary development, that of frag-
mentation of earlier rural communities; above, pp.224–26.
Page 315
Ordernr. 061297 P1
300 Wendy Davies
Orri Vésteinsson is the most explicit about his approach to reconstruction,

although not alone in his attitudes; he uses place-names, parishes and tithe areas*,
property units, assessments of natural resources, locations of churches and
chapels*, and awareness of proximity and distance, warning us to be wary of the
settlement hierarchies of prehistoric archaeology and of mechanistic models.
Those who work on England and Spain use comparable criteria, adding the
location of cemeteries and meeting places, and numbers of churches. Guy Halsall
is as systematic, but much more selective, arguing for the critical importance of the
single diagnostic criterion of burial rite in the region of Metz — or rather for
furnished burial — as the most important manifestation of community there in
the early Merovingian period, an assessment of the importance of burial rite that
is shared by other contributors. However, as time passed from sixth to seventh and
eighth centuries, he argues, societies fragmented and new powerful families emerged.
By contrast with these reconstructions, the explicitly evidenced community of the
monastic household or familia* is elaborated, in different ways, by Antonio Sennis
and Paul Fouracre. These were clearly a different kind of community from that
of the village but they equally clearly rapidly developed a capacity for corporate
action. Sennis reminds us that joining a monastery often meant changing commu-
nity, in most cases a self-conscious decision to move from one kind of group to
another, accepting a new head and new principles of association.13 Fouracre points
out that, as in the case of Marmoutier in central France, the monastic familia
could be extremely large and stretch very widely — indeed, up to 150 km. This
was a very self-conscious association: people joined it willingly and knowingly,
even if it might seem to have a detrimental effect upon their status.
One important lesson from all this is that the reconstructions are credible;
they make sense, and in making sense they tend to support the argument that
there must have been communities of association that are not directly evidenced
(an isolated farmstead simply could not survive in a country like Iceland unless it
was part of a functioning association14); however, it does not say much about the
strength of community identity nor the chronology of community development.
Another lesson is that the association of the familia or household could be much
more extensive than we might initially think; it could be a network rather than a
residential group, a network that overlapped and crosscut other communities.
13
See above, p.278.
14
See above, p.89.
Page 316
Ordernr. 061297 P1
Extent, scale and distance are considered at some length here, touching on the
issues of territoriality that feature so strongly in earlier studies.15 Communities
could stretch over a wide territory — we have just noticed a very wide monastic
familia. One kind of community could overlap another; and smaller communities
could be nested within larger units. It is rarely appropriate to think in terms of a
patchwork of totally discrete units. The concept of the supralocality*, articulated
by Julio Escalona in the context of Castile, is picked up and applied by several, in
different ways. The main argument here is that large units (50–70 km 2, with some
much larger variations), usually comprising several communities, are detectable
in the landscapes of England and Spain, of the seventh and eighth centuries in the
former region and of the tenth century (but already ancient) in the latter. Bassett,
Escalona and Martín Viso all discuss this in detail, but although the single concept
is applicable the contexts are different. Bassett’s supralocality is the large parish of
the early English minster*, subsequently to split into the parishes we know today.
Escalona’s is the administrative unit used by counts for governmental purposes.
Martín Viso’s is related to the latter but has an earlier dimension: he writes about
supralocal systems which were the spatial framework for (often reoccupied)
hilltop sites*; he goes on to detail the transformation of such supralocalities in the
tenth century when they started to be used by the counts, although they were
already an arena in which local power was expressed. How far were these large
units of England and Spain truly comparable? Do our very similar methods of
detection impose a spurious similarity? They look similar in that, as far as we can
see, they primarily had an organizational, political function, be it ecclesiastical or
secular, and they seem to have been imposed from outside (although Steven
Bassett would argue that the area which a minster served as its parish originated
in a territory, formed by the local farming population in the late fifth and sixth
centuries, which directly related to the natural landscape and its agricultural and
other economic potential; and Julio Escalona argues for ‘pre-existing building
blocks’ incorporated in different ways into new administrative systems).16 The
concept of the supralocality is clearly useful, and it is also transportable. The
important historical questions are to do with the extent to which it is anything
other than an administrative device. Did it in itself constitute community of any
sort, or was it simply a useful grouping of communities, or was it a framework
within which different communities subsequently developed?
15
16
Cf. Wickham, Framing, pp.400–04 pointing to the unusual ‘building block’ character of
some estates in western Neustria.
Page 317
Ordernr. 061297 P1
302 Wendy Davies
Icelandic patterns initially look somewhat different, although the geographical

extent of regular personal contacts may be surprising in what is often a physically
harsh landscape. We should remember, however, the relative ease of travel over
frozen landscapes.17 Each basic community, whose members were frequently
separated by at least 1 km between farmsteads, and sometimes up to 4 km, might
cover 10–20 km2 , a kind of territoriality that is difficult to match elsewhere, even
though in scale it begins to be comparable to the supralocality. Chris Callow
shows the circumstances in which people normally met and shows how people
might travel longer distances from valley to valley in the course of normal social
relationships — normal relationships, as in other parts of Europe, often being
revealed more through disputes than day-to-day functions.18 He also shows how
widely the alliance networks of richer farmers, þingmenn allied to their goðar,
stretched: none of them lived in the same valley as his chieftain; networks might
characteristically cover 50 or 60 km and in some cases of dispute protagonists
might seek support from hundreds of kilometres away, in a wider ‘political’
community. The Marmoutier monastic network discussed by Paul Fouracre is
comparable, not least in distance spanned, although it was obviously different in
function. The distance between rural manors and towns makes a related point;
although this is in essence a point about economic relationships rather than
community membership, it reminds us of further networks, and overlapping
networks; at least some of the members of those rural manors would have shared
interests with some townspeople.19 Lordly networks of patronage and clientship,
also a notable feature in Castile, provide further examples. What is interesting
here is the sense of membership, and of the benefits brought by membership, of
networks that clearly overlapped other kinds of community: people did not have
to choose one or the other; they could belong to both, or several. What is also in-
teresting in the Marmoutier case is that here serfs explicitly formed a community,
albeit a very special kind of community, unlike the excluded serfs of ninth-century
Breton and tenth-century Castilian villages (in the latter the only full members
of the community were heads of households, households which themselves often
stretched beyond the nuclear family to include others such as more distant family
members and serfs).20 Networks were necessarily selective, strengthening the links
17
See above, p.99.
18
See above, pp.75–78. Cf. Pohl, ‘Construction of Communities’, p.6.
19
20
As time passed and lordly domination increased, the distinction between free and serf
tended to dissolve but an element of this distinction remained, as is revealed in disputes over
Page 318
Ordernr. 061297 P1
between special kinds of people, but they were nevertheless communities. Defin-
ing community in ways that included some people and excluded others appears
to have been extremely common, even in smaller neighbourhoods; perhaps it was
a norm. In Iceland the figure of the outsider was acutely drawn: the outsider had
no rights to burial or community support, a distinction which very simply ex-
presses the great importance of being an insider.
In the medieval regions which are the subject of this book, the simple discrete
community seems to have been a rarity, just as the group that formed solely for
purposes of survival was a rarity; overlapping communities, wherever one looks,
were much more common, as the dominant city-based identities of Roman com-
munities concealed a host of more local group identities and the Anglo-Saxon
‘burghal’ community was an administrative imposition across pre-existing com-
munities.21 Administrative communities might come and go, but they were not
the whole story. That being the case, some interesting questions arise about the
relationship between different kinds of community. For most people, was there
a community beyond the residential household or farmstead that was in some
sense the core community, with collective identity and a capacity for collective act-
ion, the primary unit of social organization, membership of which conferred social
rights and gave access to mechanisms for redress of wrongs (like the village as it
emerged in much of continental Europe, or the group of farmers in Iceland)?
Were the networks of association subsidiary to such core groups? Which com-
munity empowered its members most? For the free individual, was it the core,
face-to-face group, or was it the complex of overlapping communities that gave
the greatest rights of participation? And how did the economic community relate
to the others?
Many of the discussions of scale in this book invoke the notion of central place
or central settlement.22 The kinds of central place vary, from the local focal point
whether a lord’s dependents should be admitted to full community rights; see Escalona above,
p.158.
21
See above, pp.200–07, 237–40, 254.
22
Cf., in a different theoretical framework, the increase in central places in southern Scan-
dinavia in the early Middle Ages: U. Näsman, ‘The Justinianic Era of South Scandinavia: An
Archaeological View’, in The Sixth Century: Production, Distribution and Demand, ed. R . Hodges
and W. Bowden (Leiden, 1998), pp.255–78 (pp.269–71), and U. Näsman, ‘Exchange and
Politics: The Eighth–Early Ninth Century in Denmark’, in The Long Eighth Century, ed. I. L.
Hansen and C. Wickham (Leiden, 2000), pp.35–68 (pp.65–68). For the classic theory, see
W. Christaller, ‘How I Discovered the Theory of Central Places: A Report About the Origin of
Page 319
Ordernr. 061297 P1
304 Wendy Davies
of the small-scale community (like a church) through the hilltop site associated
with political control of the supralocality or the city at the centre of a civitas* to
the meeting place or burial place of scattered communities, but they have at least
a notional, and often a physical, centrality in common. Firstly, as Sennis and
Fouracre make clear, the monastery was in more senses than one a central place,
be it central to a restricted locality or to a vast network: those who wrote about it
thought of the monastery as central to the surrounding physical world and the
centre of the spiritual community; it was also central in terms of authority — the
point from which orders were given, and from which collectors and inspectors
went out, a controlling point for its dependencies. Some churches were also cen-
tral places, be they the English minster church, the Icelandic chief farm church or
the Castilian village church, although not all churches were central places, witness
the many churches of Icelandic, English and Spanish countrysides. The central
churches were, again, places of control and collection, but they were also places
that were increasingly visited by their associated communities of the parish or
proto-parish — the control place in effect became a meeting place. In England
royal collection centres feature strongly too, perhaps reflecting the dominant role
of kings in the English historiography; and in all parts there is the lord’s house or
the castle or the evolving villa*-centre as control and/or collection point. In Spain
the centres of the supralocal units of alfoces*, and the hilltop sites, are notable
central places, and Escalona would see his larger ‘central settlements’ as small
capitals, seats of power, dominating the surrounding hinterland; such central
places may often have been points of political control but they could perform dif-
ferent functions, some imposed from outside, some reflecting an internal need.23
In Iceland, and perhaps in England, there were meeting places that were also a
kind of central place, not a place imposed from outside but the natural places for
people who needed to come together to do so; despite the often over-stated
ideology of þing* or folkmoot in the modern historiography, where they existed
(and some clearly did) they must have been the focal points of wider, political
and/or economic communities.24
There are clearly different kinds of central place here, from the straightforward
point of domination to the place to which people voluntarily travelled. It is
Central Places’, in Man, Space, and Environment: Concepts in Contemporary Human Geography,
ed. P. W. English and R . C. Mayfield (London, 1972), pp.601–10; P. Haggett, A. D. Cliff and
A. Frey, Locational Models (London, 1977), pp.143–45, 240–41.
23
See above, pp.159, 179–85.
24
Page 320
therefore irrelevant to ask if the central place was an expression of community or

a point of external domination — it was both and there is no necessary tension
between the two. In the literature there is a strand that argues that pre-existing
central places must have served a community function, rather like the model of
the Iron Age hill fort as refuge,25 but this is difficult to investigate. More fruitful,
and the larger question, is to think about churches. Where churches are focal, as
they certainly became as parishes developed, which came first, church or commu-
nity? Did the positioning of a church, and the need for people to attend, in effect
create community, as some contributors to this book imply, or were churches
placed where communities already existed, as others maintain? In some parts of
Europe churches can hardly have been placed where there were existing communi-
ties, for the huge parishes of Cumbria and Scandinavia, and the small parishes of
towns, relate more to the crude numbers of people to be served than to whether
or not there was a local community. Birna Lárusdóttir notes how far the
inhabitants of Reykjahverfi had to travel to go to church.26
The issue of whether the placing of a church was a stimulus or a reaction takes
us to the determinants of community identity. For some contributors the parish
is community, even if only one of several communities, and the imperative of
pastoral care, and its territorial designation, was clearly sometimes a major deter-
minant (although Paul Fouracre warns us against the ‘imagined community’ of
the parish). There were other communities, created by other kinds of stimulus,
like the hreppur* in Iceland, the community of twenty or more farming house-
holds that provided for poor relief;27 or the community of tax-paying tenants in
Frankish Gaul.28 Here there were external demands for provision of pastoral care,
or poor relief or tax. Others may have been more organic, conditioned by the eco-
nomic imperative to sustain life, by the relationships of production, or by
proximity, like the hypothetical ‘genuine’ community of Tönnies.29 Purely geo-
graphical determinants are less discussed in this book, although most contributors
think them a fundamental, if not a solely determining, element — physical
25
Cf. Glan Jones’s refuge points in his multiple estate model, e.g. his ‘The Multiple Estate as
a Model Framework for Tracing Early Stages in the Evolution of Rural Settlement’, in L’Habitat
et les paysages ruraux d’Europe, ed. F. Dussart (Liège, 1971), pp.251–67.
26
See above, p.50.
27
See Vésteinsson and Callow above, pp.102, 79–82.
28
See Halsall above, pp.227–29.
29
See above, p.5.
Page 321
Ordernr. 061297 P1
306 Wendy Davies
features, resources, agricultural potential, communications.30 Orri Vésteinsson

argues strongly against settlement form itself as a determinant: conglomerations
did not have to coincide with fiscal or property units, and often did not do so;
community identity could and did develop without nucleation of settlement.
Landlordship too might shape communities, as Frankish material in particular
seems to suggest, communities of tenants constituting the most visible of group-
ings in the seventh and eighth centuries, although there are issues here about the
extension of lordly domination over once-free proprietors.31 Although it is im-
plicit in some of the English material, change and development is not much
discussed outside these Frankish contexts; Fouracre, however, presents a clear
progression from the community of tenants of a lordly estate to an emergent
village, as the power of the dominant monastic lord declined in a world of com-
peting lordships, more cash and greater mobility. But the emerging lordships of
tenth-century Castile impacted upon existing rural communities in northern
Spain, although in parts of León and Castile these developing lordships can be
shown to have added impetus to the expression of community identity.32 While
it is entirely reasonable to postulate development from tenant community to one
or more villages in some parts, this model will not work for all parts. Lordship
impacted upon rural groups in different ways.
So what about the contrasts and the comparabilities? Despite some differences
in the chronology of development, and despite the stark differences in physical
environment, there are perhaps more similarities between Iceland, southern
England, northern Francia and northern Spain than might have been imagined.
There are clearly differences, and especially differences in settlement form, but we
have argued that these did not determine the fact or nature of community. There
is in fact a sense in which the harsher the environment, the greater the distance
the core community tended to cover and the farther selective networks extended
— ‘the ties that bound people together [. . .] outweighed the physical distances
that separated their settlements’.33 Environmental factors are invariably import-
ant but are rarely the only significant determinants, and these societies were much
30
Cf. Lárusdóttir above, pp.48–49.
31
See Halsall and Fouracre above, pp.230, 262–64. Cf. Wickham, Framing, pp.398–406, on
the Paris basin.
32
See W. Davies, ‘The Early Middle Ages and Spanish Identity’, in Power and Identity: Essays
Presented to Rees Davies, ed. H. Pryce and J. Watts (Oxford, forthcoming).
33
Above, p.112.
Page 322
Ordernr. 061297 P1
too complex to associate purely in order to survive. As for the government/locality

relationship, at present the evidence that governments, lords and churches had a
hand in creating communities looks as strong as that which suggests that they
brought out of obscurity existing local groupings, not least because groups fre-
quently changed, forming, combining, splitting and re-forming.
We did not find much evidence of physical markers left by local communities,
although they were widely used to mark properties. We found communities and
associations of many kinds, from the hamlet to the far-flung network, overlapping
and nested; rarely did we find the discrete community. We hope we have provided
plenty of methodological guidance — particularly on how to spot the community
that does not leave much of a mark in surviving contemporary sources. We
answered some questions but posed more. The important questions that remain
are to do with the formation of ‘core’ communities beyond the residential house-
hold, their capacity for collective action and their relationship with the several
networks of association to which some (or all) of their members also belonged.
Photo 14.2. The group at the hillfort of La Cerca, Quintanar de la Sierra.
Page 323
Ordernr. 061297 P1
Page 324
Ordernr. 061297 P1
G LOSSARY
compiled by Guy Halsall
NOTE An asterisk following a word indicates that the word is a main entry
elsewhere in the Glossary.
Alfoz (pl. alfoces). Castilian. The standard term — but not the only one — for a
medieval Castilian administrative district. For the Castilian counts (up to 1037)
and kings (thereafter), the alfoces were the territorial framework for both gov-
ernance and patrimony management. The free population in each of those units
were subject to jurisdiction, probably work renders, and to military obligations
which were later converted into taxes. The original network of Castilian alfoces
seems to have been created in the late ninth to early tenth centuries on the basis
of existing territorial units of supralocal* kind.
Alþing. Icelandic. Iceland’s central or ‘national’ assembly, held annually in the

summer at Þingvellir in south-west Iceland. Presided over by a Lawspeaker, the
alþing was the venue for the Law court and two tiers of judicial courts. Goðorðs-
menn* held the votes in the Law Court and had the right to appoint judges to the
judicial courts but not sit on the panels themselves.
Ancilla (pl. ancillae). Latin. A female serf. The term has connotations of domestic
service, but usually appears to denote female servitude in general.
Ango (pl. angones). Latin? A heavy iron javelin with a barbed head, like a harpoon,
found in lavishly furnished sixth-century male burials in the Frankish territories
and elsewhere, and described in use by Procopius.
Page 325
Ordernr. 061297 P1
310 Guy Halsall
Bænhús (pl. bænhús). Icelandic. Literally ‘prayer-house’. Chapel*. The lowest

ranking type of church, serviced by a priest from a nearby parish* church
(sóknarkirkja*) twelve times a year or less. As a rule bænhús had cemeteries
associated with them.
Bær (pl. bæir). Icelandic. Farmstead, general term referring to the site of a farm. See
also lögbýli* and hundrað*.
Banal lord. A lord who exercised a range of judicial, fiscal and military rights,
many of which had originally been derived from royal or public authority.
Characteristic of Frankish areas in the central Middle Ages.
Burh. Old English. A fort. Specifically the term is used to describe the royal
fortresses constructed — or placed into an administrative system — in England
during the reign of Alfred the Great and his successors.
Bygð (pl. bygðir). Faroese. Literally, ‘settlement’. Hamlet or village. A geograph-

ically discrete settlement including one or more fyrndarbýlingur*.
Byggð (pl. byggðir). Icelandic. Literally, ‘settlement’. Neighbourhood. Less specific

and generally more extensive than hverfi*, þorp* or Faroese bygð*.
Cartulary. A book into which charters and documents from an archive were
copied. Theoretically lacking legal probatory status, cartularies were normally
intended for easy archival reference and consultation. Following the loss or
destruction of original charters, cartularies are often our only source for early
medieval charter texts.
Castellum (pl. castella). Latin. Originally a small fort and, in the later Roman
period, a fortified settlement of lower status than civitates* or municipia* . By the
tenth century, a fortification that might also be a residence, or part of a settle-
ment, unlike the ‘tower’ which seems to have been purely a garrisoned defensive
structure.
Chapel. A church which lacked independent parish* status. See also bænhús*,
parochial chapel*.
Chapelry. The defined area served by a parochial chapel*.
Page 326
Ordernr. 061297 P1
GLOSSARY 311
Civitas (pl. civitates). Latin. During the Roman Empire, an administrative unit,
in Britain and Gaul often preserving the territory of a pre-Roman tribal unit,
though in Spain with a more complex relationship to pre-Roman units, and
usually based around a city, for which the same term — civitas — is used. The
civitas was the building block of many western imperial provinces and, following
the conversion of Constantine, often became a Christian diocese. In early
medieval sources the term preserves its meaning as an area around a major
administrative centre but by the eleventh century it is often used to mean ‘town’.
Cognatio (pl. cognationes). Latin. In Roman inscriptions, this term is used to mean
relationship by birth. Kin groupings in the Celtiberic language are normally
expressed by an adjective in the genitive plural deriving from a personal name.
Colibertus (pl. coliberti). Latin. A type of serf. How coliberti differed from servi*
is not made clear in documents from north-west France in the eleventh century.
Conventus iuridicus. Latin. In the Roman period, a large jurisdictional district

whose population depended on a capital for the administration of justice. The
conventus seems to have gone out of use during the third century AD.
Curacy. The office of a priest, often very poorly paid, who assisted a rector or
vicar, that is, of an assistant priest (as opposed to a parish* priest).
Curia. Latin. In the Roman period, the assembly of the (urban) local Senate.
Demesne land. The parts of an estate or (in England) manor, which its lord kept
in his own hands for direct management in contrast to those parts leased out to
tenants.
Denarius (pl. denarii). Latin. Originally an early Roman coin, by the later part of
the European early Middle Ages, it was a silver coin: the ‘penny’. Conventionally
there were twelve denarii to the solidus*.
Emir. From the Arabic ’amir. In the Iberian regions under Muslim rule, the title
of the highest ruler from the conquest to 929 AD, when Abd al-Rahman III took
the title of Caliph.
Page 327
Ordernr. 061297 P1
312 Guy Halsall
Familia. Latin. Household; group of dependants of a lord or of a church or of a

monastery. Often associated with an ensemble of serfs on an estate, in the case of
a monastery the group could include the monks as well as better-off clients, tribu-
taries and serfs, for all were pledged to serve forever.
Feorm. Old English. The tribute paid to an Anglo-Saxon king, normally in the
form of a wide range of foodstuffs, and normally in an amount determined by the
number of hides* at which a payee’s land was assessed.
Fines. Latin. Literally ‘limits’. In early medieval Frankish charters, the term used
for a small geographical territory. Sometimes fines are associated with particular
settlements, but often it seems that settlement within them was dispersed. Fines
often have place-names ending with -acum or -ensis, both implying a region. The
Germanic equivalent of this term is marca*.
Formula. Latin. A standard phrase in a document, or a model document.
Francisca (pl. fransciscae). Latin? A throwing axe commonly found in late fifth-
and sixth-century cemeteries in Frankish territories and described in use by
Sidonius Apollinaris and Agathias. The francisca might have been developed in
the very late Roman army, rather than, as usually assumed (but without evidence)
being a Frankish import. Isidore of Seville believed that the Franks took their
name from the francisca.
Fyrndarbýlingur (pl. fyrndarbýlingar). Faroese. Literally, ‘ancient-farm’. Assessed

farm, one or more of which made up a bygð*. Fyrndarbýlingar could be subdivided
into numerous households.
Glebe land. Land with which a late medieval English church was endowed, and
from which its priest derived most of his subsistence.
Goðorðsmaðr (pl. goðorðsmenn). Icelandic. The owner of a goðorð, a seat in the Law
court (the legislature) of the alþing*, the national assembly. Normally associated
with temporal power, but although as a group goðorðsmenn represented the
topmost tier of höfðingjar* in Iceland, as individuals they could be powerless
placeholders.
Page 328
Ordernr. 061297 P1
GLOSSARY 313
Hide. A unit of assessment of liability to pay the public burdens which were
imposed on most land in Anglo-Saxon England.
Hilltop sites. The late antique and early medieval periods saw a flourishing of
(normally fortified) hilltop settlements, sometimes reoccupying deserted Late
Iron Age hill forts. Hilltop sites varied much in size and functions. Many were
created by local communities (often in relation to a residential area nearby), but
some were set up by central authorities to serve as central places for territorial
control.
Hoba (pl. hobae). In early medieval Frankish charters, the latinized version of the
Germanic word for mansus*. In the charter evidence there seems to be no
technical difference between the two terms.
Höfðingi (pl. höfðingjar). Icelandic. Chieftain. General term for a significant or de

facto leader, but who might not be a goðorðsmaðr*.
Hospitium. Latin. In the Roman period, a formal guest-friendship and patronage

relationship linking individuals and/or communities; both parties established
privileged links between them by granting each other especial rights and duties.
Hreppur (pl. hreppar). Icelandic. Commune or district. An association of twenty

or more lögbýli* forming a geographically continuous area, collectively responsible
for poor-relief and insurance in case of fire. Could also coordinate pest control
and communal grazing of livestock. As a rule hreppar were larger units than
sóknir*.
Hundrað (pl. hundruð). Icelandic. Literally, ‘one hundred’. Unit of valuation of

landed property, traditionally regarded as equivalent to 120 ells of homespun
cloth or one cow or six ewes with lambs. Lögbýli* were assessed in hundruð. An
average Icelandic lögbýli was valued at 20–24 hundruð while a 60-hundruð lögbýli
would be considered a stórbýli — a major farm or estate.
Hundred. An English administrative land unit, focussed on a court to which

individual constituent manors sent representatives; it was a formal subdivision of
a shire* (known as a wapentake in much of northern and eastern England).
Page 329
Ordernr. 061297 P1
314 Guy Halsall
Hverfi (pl. hverfi). Icelandic. Neighbourhood. A cluster of bæir* or the bæir of a

geographically discrete area. See also byggð* and þorp*.
Infanzón (pl. infanzones). Castilian. In eleventh-century Castile, the lesser rank

of nobility. Though etymologically referring to nobility by birth, it was seen fit to
translate the Latin milites*. Infanzones are often recorded as forming the lesser
ranks in aristocratic clientage networks, but also acting collectively as a body of
local notables in defence of the rights of local/supralocal* communities.
King’s tûn. Old English. An Anglo-Saxon royal estate and its chief buildings, the
latter being the regular collecting-point for the feorm* of the inhabitants of the
surrounding district. (It is usually referred to as cyninges tûn in Old English
sources and as villa* regalis in Latin ones.)
Lögbýli (pl. lögbýli). Icelandic. Literally, ‘legal-farm’. Assessed farm, a taxable and
tradable unit of land with at least one farming household. Communal obligations
were levied on the lögbýli and not on its subdivisions. See also bær*.
Mansus (pl. mansi). Latin. Farm. Mansi were the constituent units of Merovin-
gian villae*. Sometimes the word refers to dependent tenancies within a landhold-
ing, but it can also refer to an apparently extensive estate with dependencies and
appurtenances, and be described in exactly analogous terms to unitary villa-
estates*.
Marca. Old High German. In early medieval Frankish charters, the Germanic
equivalent of fines*. There seems to have been no technical difference between the
two terms.
Merindades menores. Castilian. Administrative districts, each under the rule of

one merino (royal officer), which were developed between the late twelfth and
early thirteenth century. The newly created network superseded the alfoz* system.
The merindades menores were substantial territories, each comprising several of
the old alfoces. In the thirteenth century, there was yet a higher level, the Merin-
dad Mayor de Castilla, roughly coterminous with the county of Old Castile.
Milites. In classical Latin, the term means simply ‘soldiers’, but by the eleventh
century it had come to denote mounted soldiers, and as the latter rose in social
status, it became a general term for the lesser ‘nobility’.
Page 330
Ordernr. 061297 P1
GLOSSARY 315
Minster. A complex ecclesiastical settlement headed by an abbess, abbot, or man

in priest’s orders; which contains nuns, monks, priests or laity in a variety of
possible combinations, and is united to a greater or lesser extent by their liturgy
and devotions; which may perform or supervise pastoral care to the laity, perhaps
receiving dues and exerting parochial authority; and which may sometimes act as
a bishop’s seat, while not depending for its existence or importance on that
function [with grateful thanks to John Blair for this definition].
Mother-church. In England, a church with pastoral responsibility for an area

much larger than a standard late medieval/modern parish*, and with supervisory
powers over other, lesser public churches (i.e. chapels*) in that area.
Mozarabic. From the Arabic musta‘rib, denoting ‘arabized’. Used by a prominent

historiographical tradition to designate the Christian population living under
Muslim rule in Iberia, and also a brand of pre-Romanesque style art and archi-
tecture (also manuscript illumination) mixing a strong Visigothic background
together with Arabic aesthetic and cultural influence.
Municipium (pl. municipia). Latin. In the Roman period, a town with a formal
right to self-governance, whose free inhabitants were in possession of the status
of Roman citizenship.
Oppidum (pl. oppida). Latin. Although it occurs in texts from the Roman period,
in the Duero basin (and elsewhere) this term is usually employed by modern
historians to designate the largest late pre-Roman Iron Age (often referred to as
Celtiberic) settlements. These were substantial sites, often provided with walls or
ramparts and some form of urbanism, as opposed to the more elementary hill
forts. Some oppida became Roman towns and eventually achieved a municipal
status as municipia*.
Parish. The geographical area inhabited by the people who received pastoral care
from a church with designated responsibility for providing care — either directly
or through the intermediacy of another lesser church (alias a chapel*), which was
ecclesiastically subject to the responsible church. See also sókn*, tithe area*.
Parochial chapel. A public church which, although lacking independent parish*

status, assisted an English mother-church* in the delivery of pastoral care to a
substantial part of the latter’s parish.
Page 331
Ordernr. 061297 P1
316 Guy Halsall
Patronus. Latin. In the Roman period, a prominent individual acting as a protector

(defender) of the interests of individuals, whole cities or even entire provinces. See
further tabula patronatus*.
Placitum (pl. placita). A formal court hearing, or the document recording such a
hearing, or a legal case that might be protracted, that is, involving several hearings
of various kinds. In eleventh-century documents from north-west France the term
most often refers to a formal court hearing.
Polyptychs. Literally meaning ‘many-leaved’ documents, these were extensive

surveys of church estates, produced in the ninth and tenth centuries, originally at
the bidding of the Carolingian rulers of the Franks.
Praeceptum (pl. praecepta). A royal charter which, at least in theory, royal

authority made binding.
Realengo. Castilian. In medieval Castile, the king’s specific lordship, one of four
main seigneurial forms which were recognized in the central medieval period,
together with abadengo (ecclesiastical lordship), solariego (secular aristocratic
lordship) and behetría (a complex, archaic hybrid of royal and aristocratic lord-
ship). It notionally derived from royal control of (especially military) obligations
owed by supralocal units* directly subjected to the king with exclusion of other
lordly interests.
Ringwork. In England, a private fortification, usually approximately circular,

consisting of a bank and ditch and often with a strongly defended entrance.
Sceatta. From the Old English word sceatt (‘wealth’), in some Old English sources
it is also used for a coin, 1/240 to 1/250 of a pound and thus roughly equivalent
to Frankish denarii*. Modern numismatists, historians and archaeologists have
applied the word to specific series of middle Saxon silver coins introduced from
the seventh century.
Servus (pl. servi). Latin. In classical Latin, a slave, but the meaning of the term
evolves with the institution of servitude. By the eleventh century the servus had
become a serf, that is, a peasant dependent on a lord, with restrictions on marriage
and inheritance, the condition being hereditary. In West Frankish sources servi
acknowledged their unfree status by placing four pennies on their heads.
Page 332
Ordernr. 061297 P1
GLOSSARY 317
Shieling (sel in Icelandic). A temporary dwelling place in the outfield where milk
cows and ewes were grazed and milked during the summer. Sometimes shielings
became permanent farms.
Shire. Originally meaning a ‘share’ of land, the shire system formed the basis of
provincial organization in the Late Anglo-Saxon period in England. These large
territories were divided into hundreds*.
Sókn (pl. sóknir). Icelandic. Literally, ‘attendance’. Originally the area for which
services were provided by a particular church, it was later a territorially defined
parish* as conveyed by the word þing*.
Sóknarkirkja (pl. sóknarkirkjur). Icelandic. Literally, ‘attended church’. Parish*

church. A church with a resident priest who might also service útkirkjur which
had their own tithe areas*.
Solidus (pl. solidi). Latin. Originally, from the fourth century, a gold coin weigh-
ing 1/72 of a pound. In medieval western Europe, after the end of gold coinage,
it came to refer to weights of metals rather than to coins themselves. It was usually
reckoned that one solidus was worth twelve silver denarii*.
Suburbium. A word of Roman origin used, in tenth- and eleventh-century Castile,

to designate a central settlement’s hinterland, normally coincident with an admin-
istrative district, as an alternative to the more usual alfoz*. It was used, without
distinction, for proto-urban and clearly non-urban hierarchical settlements.
Supralocal unit. In the early medieval Castilian landscape there are found terri-
torial blocks which included several settlements (of village and hamlet size), with
both individual and common use of landed resources (especially valleys) within
the block; they were often in the order of 50–100 km 2, but could be as large as 250
km2. In origin they relate to the exploitation of the land, and to the organization
of space, by local communities; some have central places, such as hilltop sites*.
Some of them were used by Castilian rulers in the tenth century and later in order
to build a structure through which to exercise political control, especially through
the network of alfoces*.
Page 333
Ordernr. 061297 P1
318 Guy Halsall
Tabula patronatus. Latin. In the Roman period, a legal document (often inscribed
on a metal tablet) by which a patronus* granted his protection to an individual,
group or community.
Tephra. Volcanic ash. Layers of volcanic ash attributed to known eruptions are an
important dating tool in Icelandic archaeology. The Veiðivötn tephra of 871±2
(‘Landnám tephra’) and the Hekla tephra of 1104 are important chronozones for
the colonization period in Iceland.
Tessera (pl. tesserae). Latin. In the Roman period, small and light metallic or
wooden pieces intended to serve as proof of identity for either an individual (with
the purpose of distribution of silver or wheat, access to the games, etc.) or the
parties involved in a pact of hospitality. Tesserae hospitalis are pairs of bronze
pieces (often cut in the shape of animals or objects) split in two symmetrical parts,
each one inscribed on the smoother face with the text of a pact of hospitality. In
Roman Spain, the text of these tesserae is written in both Celtiberic and Latin.
Tithe area. An area, within which all the lögbýli* paid tithe to a particular church.
In the Icelandic parish* structure, útkirkjur (‘out churches’ dependent upon par-
ish churches — sóknarkirkjur*), but not bænhús*, had their own tithe areas, so that
a þing* could be made up of many tithe areas, some consisting of only one lögbýli.
Þing (pl.). Icelandic. The area served by a priest. Could include a number of
churches of different types (e.g. bænhús*) although as a rule þing only included one
church where the priest had his abode (sóknarkirkja*). In the singular (also þing)
this word means ‘assembly’.
Þorp (pl. þorp). Icelandic. Neighbourhood or hamlet. Cluster of bæir* , normally

more clustered than either hverfi* or byggðir*.
Vicus (pl. vici). Latin. In the Roman period, a rural agglomeration of lower status
than municipia* or civitates*. When fortified, in Gaul generally in the Late Roman
period, these could be referred to as castra or castella*. In tenth-century north-
west France, the term is used to mean a rural settlement without a resident lord,
possibly inhabited by peasants of both free and unfree status. The term occurs
occasionally in early medieval Spanish texts, often designating the centres of
alfoces*.
Page 334
Ordernr. 061297 P1
GLOSSARY 319
Villa. Latin. This word changed its meaning during the early medieval period, in
accordance with chronological and regional changes and variations in the rural
settlement pattern. In the Roman period it meant the physical buildings at the
centre of a villa-estate*. In the immediately post-imperial west, it was generally
used to mean a small territory, sometimes a single estate, but more often an area
within which a number of people could hold lands. In tenth- and eleventh-
century Castilian charters this is the word usually employed to designate a rural
settlement of village rank and/or its territory. In eleventh-century West Frankish
sources the term appears to denote a type of settlement associated with a lord, or
lord’s residence, although several lords might have a share in a single villa.
Villa-estate. In the Roman period, an estate worked by a dependent population,

comprising land, a central building (the villa-building or pars residentialis),
production and storage facilities, and other dependent dwellings.
Page 335
Ordernr. 061297 P1
Page 336
Ordernr. 061297 P1
B IBLIOGRAPHY
Primary Sources
Manuscript Sources
The Árni Magnússon Institute, Reykjavík, AM 66a 8vo , Máldagakver Odds biskups Einarssonar.
Biblioteca Apostolica Vaticana, Archivio di S. Pietro, Caps. 72, fasc. 53.
London, British Library, MS Harley 595
Paris, Bibliothèque Nationale de France, fonds latin 5411, ‘Liber instrumentorum de possessio-
nibus, rebus, sive dignitatibus, quas Casauriense monasterium habuit, habet, vel habere debet’.
Printed Sources
Das älteste Traditionsbuch des Klosters Mondsee, ed. G. Rath and E. Reiter, Forschungen zur
Geschichte Oberösterreichs, 16 (Linz, 1989).
Anales Castellanos I, ed. M. Gómez Moreno (Madrid, 1917).
The Anglo-Saxon Chronicles, ed. M. Swanton (London, 2000).
Árnessýsla: Sýslu og sóknalýsingar Hins íslenzka bókmenntafélags 1839–1843 og Lýsing Ölfushrepps
anno 1703 eftir Hálfdan Jónsson, ed. S. Sigmundsson (Reykjavík, 1979).
Asser’s Life of King Alfred, ed. W. H. Stevenson(Oxford, 1904).
Bacon, J., Liber Regis, vel Thesaurus Rerum Ecclesiasticarum (London, 1786).
Baños Rodríguez, G., Corpus de Inscricións romanas de Galicia, vol. II: Provincia de Pontevedra
(Santiago, 1994).
Becerro Gótico de Cardeña, ed. L. Serrano (Valladolid, 1910).
Bede, Ecclesiastical History of the English People, ed. and trans. B. Colgrave and R . A. B. Mynors
(Oxford, 1969).
Benedicti Regula, ed. R . Hanslik, Corpus scriptorum ecclesiasticorum Latinorum, 75 (Vienna,
1960).
Beowulf, ed. J. Alexander (Harmondsworth, 1978).
Calendar of Documents Preserved in France, vol. I, ed. J. H. Round (London, 1899).
Calendar of Patent Rolls, 1292–1301 (London, 1895).
Page 337
Ordernr. 061297 P1
322 Bibliography
Calendar of the Register of Adam de Orleton Bishop of Worcester 1327–1333, ed. R . M. Haines,
Worcestershire Historical Society, n.s., 10 (London, 1979).
Cartulaire de l’Abbaye de Gorze (Mettensia 2), ed. A. d’Herbomez (Paris, 1898; introduction
dated 1902).
Cartulaire de Marmoutier pour le Dunois, ed. M. Emille Mabille (Chateaudun, 1874).
Cartulaire de Marmoutier pour le Vendômois, ed. M. de Trémault (Paris, 1893).
Cartulario del Infantado de Covarrubias, ed. L. Serrano (Valladolid, 1907).
Cartulario de San Juan de la Peña, ed. A. Ubieto Arteta (Valencia, 1962).
Cartulario de San Millán de la Cogolla (759–1076), ed. A. Ubieto Arteta (Valencia, 1976).
Cartulario de San Pedro de Arlanza (antiguo monasterio benedictino), ed. L. Serrano (Madrid,
1925).
Los Cartularios Gótico y Galicano de Santa María de Valpuesta (1090–1140), ed. S. Ruiz de
Loizaga (Vitoria, 1995).
Cartularium Saxonicum, ed. W. de G. Birch, 3 vols (London, 1885–93).
Chartularium Cupersanense, ed. D. Morea (Montecassino, 1892).
Chronica Monasterii Casinensis, ed. H. Hoffmann, MGH SS, 24 (Hannover, 1980).
Il Chronicon di Benedetto di S. Andrea del Soratte e il Libellus de imperatoria potestate in urbe
Roma, ed. G. Zucchetti (Rome, 1920).
CIL, vol. II: Inscriptiones Hispaniae, consilio et auctoritate Academiae Litterarum Regiae Borussicae,
ed. E. Hübner (Berlin, 1869; suppl. Berlin, 1892).
CIL, vol. VI.2: Inscriptiones urbis Romae latinae, consilio et auctoritate Academiae Litterarum
Regiae Borussicae, ed. E. Bormann, G. Henzen and Ch. Huelsen (Berlin, 1882).
Codex Diplomaticus Fuldensis, ed. E. F. J. Dronke (Cassel, 1850).
Codice diplomatico del monastero benedettino di S. Maria di Tremiti (1005–1237), ed. A. Petrucci
(Rome, 1960).
Colección Diplomática de los Condes de Castilla, ed. M. Zabalza Duque (Salamanca, 1998).
Colección Diplomática del Monasterio de Sahagún (siglos IX y X ), ed. J. M. Mínguez (León, 1976).
Colección Diplomática del Monasterio de Sahagún, II (1000–1073), ed. M. Herrero de la Fuente
(León, 1988).
Colección Diplomática San Salvador de Oña (822–1284), ed. J. del Álamo (Madrid, 1950).
Colección Documental del Archivo de la Catedral de León, IV (1032–1109), ed. J. M. Ruiz Asencio
(León, 1990).
Colección Documental del Monasterio de San Pedro de Cardeña, ed. G. Martínez Díez (Burgos,
1998).
Collectionis Bullarum Sacrosantae Basilicae Vaticanae (Rome, 1747).
Cronaca di Novalesa, ed. G. C. Alessio (Turin, 1982).
De mensura orbis terrae, ed. J. Tierney and L. Bieler, Scriptores Latini Hiberniae, 6 (Dublin,
1967), pp.44–103.
De miraculis s. Adalhardi, ed. O. Holder-Egger, MGH SS, 15.2 (Hannover, 1888), pp.859–65.
de Palol, P., and J. Vilella, Clunia II: La epigrafia de Clunia (Madrid, 1987).
A Descriptive Catalogue of Ancient Deeds, 6 vols (London, 1890–1915).
De situ orbis, ed. R . Quadri (Padua, 1974).
Reykjavík, 1853–1976).
Documentación de la catedral de Burgos (804–1183), ed. J. M. Garrido Garrido (Burgos, 1983).
Documentación de la catedral de Palencia (1035–1249), ed. T. Abajo Martín (Palencia, 1986).
Page 338
Ordernr. 061297 P1
Bibliography 323
Documentación del monasterio de Santo Domingo de Silos (954–1254), ed. M. C. Vivancos

(Burgos, 1988).
Domesday Book, vol. XVI: Worcestershire, ed. F. Thorn and C. Thorn (Chichester, 1982).
Domesday Book, vol. XXIII: Warwickshire, ed. J. Plaister (Chichester, 1976).
Domesday Book seu Liber Censualis Willelmi Primi Regis Angliae, ed. A. Farley, 2 vols (London,
1783).
Ecclesiastical Terriers of Warwickshire, vol. II, transcribed and ed. D. M. Barratt, Dugdale Society,
27 (1971).
Ex miraculis et translationibus s. Bavonis, ed. O. Holder-Egger, MGH SS, 15.2 (Hannover, 1888),
pp.589–99.
Eyfirðinga sögur, ed. J. Kristjánsson, Íslenzk fornrit, 9 (Reykjavík, 1956).
Formulae Turonenses, in Formulae Merowingici et Karolini Aevi, ed. K. Zeumer, MGH Leges in
quarto, Sectio 5 (Hannover, 1886).
The Fourth Book of the Chronicle of Fredegar with its Continuations, ed. and trans. J. M. Wallace-
Hadrill (London, 1960).
Die Gesetze der Angelsachsen, ed. F. Liebermann (Halle, 1903).
Grágás. Elzta lögbók Íslendinga. Útgefin eptir skinbókinni í bókasafni konungs, 2 vols
(Copenhagen, 1852; repr., Odense, 1974).
Gregory of Tours, Histories, ed. B. Krusch and W. Levison, MGH SRM, 1.1 (Hannover, 1951).
———, Vitae Patrum, ed. B. Krusch, MGH SRM, 1.2 (Hannover, 1885).
Hagskinna: Sögulegar hagtölur um Ísland, ed. G. Jónsson and M. S. Magnússon (Reykjavík, 1997).
Hydatius, Chronicon, ed. J. Campos (Salamanca, 1984).
Ibn Idari, Historia de Al-Andalus, ed. F. Fernández González (Málaga, 1999).
‘Instrumenta ad tomum xi’, in Gallia Christiana in Provincias Ecclesiasticas Distributa vol. XI, ed.
P. Piolin (Paris, 1874).
Ionae Vitae sanctorum Columbani, Vedastis, Iohannis, ed. B. Krusch, MGH SRG in usum
scholarum, 37 (Hannover, 1905).
Jarðabók Árna Magnússonar og Páls Vídalíns, 13 vols (Copenhagen, 1913–90).
Die Konzilien der karolingischen Teilreiche, 843–859, ed. W. Hartmann, MGH Concilia, 3
(Hannover, 1989).
Landnámabók, ed. J. Benediktsson, Íslenzk fornrit, 1 (Reykjavík, 1968).
Lárusson, B., The Old-Icelandic Land Registers (Lund, 1967).
The Laws of the Earliest English Kings, ed. A. J. Robertson (Cambridge, 1922).
Laws of Early Iceland: Grágás, the Codex Regius of Grágás, with Material From Other Manuscripts,
trans. A. Dennis, P. Foote and R . Perkins, 2 vols (Winnipeg, 2000).
Leges Burgundionum, ed. F. de Salis, MGH Legum Sect I, 2.1 (Hannover 1892).
Lex Ribvaria, ed. F. Beyerle and R . Buchner, MGH Legum Sect. I, 3 (Hannover, 1951).
Liber de Servis Majoris Monasterii, ed. Ch. L. Grandmaison, Publications de la Société archéolo-
gique de Touraine, 16 (Tours, 1864).
Libro Becerro de las Behetrías, ed. G. Martínez Díez (León, 1981).
Ljósvetninga saga, ed. B. Sigfússon, Íslenzk fornrit, 10 (Reykjavík, 1940).
Macrobius, Commentarium in somnium Scipionis libri duo, ed. L. Scarpa (Padua, 1981).
Manntal á Íslandi árið 1703 tekið að tilhlutan Árna Magnússonar og Páls Vídalíns (Reykjavík,
1924–47).
Martial, Epigramas, ed. M. Dolç (Barcelona, 1949–50).
Page 339
Ordernr. 061297 P1
324 Bibliography
Martín Viso, I., ‘Monasterios y Poder Aristocráticos en Castilla en el siglo XI’, Brocar, 20 (1996),
91–133.
Martinus Turonensis, Vita Alcuini abbatis Turonensis, ed. W. Arndt, MGH SS, 15.1 (Hannover,
1887), pp.182–97.
Minutes and Accounts of the Corporation of Stratford-upon-Avon and Other Records, 1553–1620,
vol. IV , transcribed R . Savage and others, Publications of the Dugdale Society, 10 (London,
1929).
Muratori, L. A., Rerum Italicarum Scriptores, 25 vols (Milan, 1726).
Nonarum Inquisitiones in Curia Scaccarii temp. Regis Edwardi III (London, 1807).
Pactus Legis Salicae, ed. K.-A. Eckhardt, MGH Legum Sect. I, 4.1 (Hannover, 1962).
Pérez de Urbel, J., Historia del Condado de Castilla, 3 vols (Madrid, 1945).
Pliny, Naturalis Historia, ed. C. Mayhoff (Leipzig, 1906).
Rabani Mauri. De universo libri viginti duo, ed. J. P. Migne, Patrologia Latina, 111 (Paris,
1841–64), cols 9–614.
Regesta Regum Anglo-Normannorum, 1066–1154, 4 vols (Oxford, 1913–69), III, ed. H. A.
Cronne and R . H. C. Davis.
The Register of the Diocese of Worcester during the Vacancy of the See, usually called ‘Registrum Sede
Vacante’, ed. J. W. Willis Bund (Oxford, 1897).
Reykdœla saga ok Víga Skutu, in Ljósvetninga saga með þáttum. Reykdœla saga ok Víga-Skútu.
Hreiðars þáttr, ed. B. Sigfússon, Íslenzkt Fornrit, 10 (Reykjavík, 1940).
Rimbert, Vita Anskarii, ed. W. Trillmich, in Quellen des 9. und 11. Jahrhunderts zur Geschichte
des Hamburgischen Kirche und des Reiches, ed. W. Trillmich and R . Buchner (Darmstadt,
1961).
(Madrid, 1943).
Sermo de relatione corporis Beati Vedasti, in Monumenta Vedastina Minora, ed. O. Holder-Egger,
MGH SS, 15 (Hanover, 1887), pp.396–405.
Sturlunga Saga, ed. J. Jóhannesson, M. Finnbogason and K. Eldjárn, 2 vols (Reykjavík, 1946).
Sturlunga Saga, ed. Õ. Thorsson (Reykjavík, 1988).
Testamentum Bertramni Cenomannis, in Diplomata, Chartae, Epistolae, Leges aliaque instrumenta
ad Res Gallo-Francicas spectantia, ed. J. M. Pardessus, 2 vols (Paris, 1843–49), I, 197–215.
Traditiones Wizenburgenses: Die Urkunden des Klosters Weissenburg, 661–864, ed. K. Glöckner
and A. Doll (Darmstadt, 1979).
Translatio corporis s. Benedicti in Franciam, ed. R . Weber, Revue bénédictine, 62 (1952), 140–42.
Translatio et miracula sanctorum Senesii et Theopontii, ed. P. E. Schramm, MGH SS, 30.2
(Leipzig, 1934), pp.984–92.
Urkundenbuch des Klosters Fulda, ed. E. E. Stengel (Marburg, 1956).
Valor Ecclesiasticus temp. Henr. VIII auctoritate regia institutus, 6 vols (London, 1810–34).
Víga-Glums saga, in Eyfirðinga sögur, ed. J. Kristjánsson, Íslenzk fornrit, 9 (Reykjavík, 1956),
pp.1–98.
Vita Rimberti, ed. G. Waitz, MGH SRG in usum scholarum, 55 (Hannover, 1884), pp.80–100.
Vita Romarici Abbatis Habandensis, ed. B. Krusch, MGH SRM, 4 (Hannover, 1902).
The Vita Wulfstani of William of Malmesbury, ed. R . R . Darlington (London, 1928).
Warwickshire Feet of Fines, vol. I, ed. F. C. Wellstood, Publications of the Dugdale Society, 11
(London, 1935).
Page 340
Ordernr. 061297 P1
Bibliography 325
Secondary Sources
Abásolo, J. A., and R . García Rozas, Carta arqueológica de la provincia de Burgos: Partido judicial
de Salas de los Infantes (Burgos, 1980).
Abásolo Álvarez, J. A., Epigrafía romana de la región de Lara de los Infantes (Burgos, 1974).
———, ‘Las estelas decoradas de la región de Lara de Los Infantes: Estudio Iconográfico’, Boletín
del Seminario de Arte y Arqueología de la Universidad de Valladolid, 43 (1977), 61–90.
———, Las vías romanas de Clunia (Burgos, 1978).
Abásolo Álvarez, J. A., and F. Pérez, ‘Excavaciones en Salinas de Rosío (Burgos)’, Noticiario
Arqueológico Hispano, 24 (1985), 159–264.
Abásolo Álvarez, J. A., and I. Ruiz Vélez, Carta arqueológica de la provincia de Burgos: Partido
judicial de Burgos (Burgos, 1977).
Albertos, Mª. L., ‘El Culto a los montes entre los Galaicos, Astures y Berones y algunas de las
deidades más significativas’, Estudios de Arqueología Alavesa, 6 (1974), 47–157.
Albertos Firmat, M. L., ‘Organizaciones suprafamiliares en la Hispania Antigua’, Studia
Archaeologica, 37 (1975), 5–66.
Alcock, L., Arthur’s Britain: History and Archaeology AD 367–634 (London, 1971).
Alföldy, G., ‘Notes sur le droit de cité et la nomenclature dans l’Empire Romain’, Latomus, 25
(1966), 37–57.
Algorreta, J. A., and others, ‘Castro de Berbeia (Barrio-Álava): Memoria de excavaciones,
Campaña de 1972’, Estudios de Arqueología Alavesa, 8 (1975), 221–92.
Álvarez Borge, I., ‘Estructuras de poder en Castilla en la alta edad media: señores, siervos, vasallos’,
in Señores, Siervos, Vasallos en la Edad Media (Pamplona, 2002), pp.269–308.
———, Monarquía feudal y organización territorial: Alfoces y merindades en Castilla (siglos
X – XIV ) (Madrid, 1993).
Anderson, A. H., Alexander’s Gate: Gog and Magog and the Inclosed Nations (Cambridge, MA,
1932).
Andrés Ordax, S., ‘Arte Burgalés de la alta edad media’, in II Jornadas Burgalesas de Historia:
Burgos en la Alta Edad Media (Burgos, 1991), pp.125–40.
Année Épigraphique (1953).
Arce, J., ‘Los gobernadores de la Diocesis Hispaniarum (ss. IV – V d.C.) y la continuidad de las
estructuras administrativas romanas en la Península Ibérica’, Antiquité Tardive, 7 (1999),
73–83.
———, El último siglo de la España Romana (274–409) (Madrid, 1986).
Arge, S. V., ‘Føroysk búsetingarsøga – tær fornfrøðiligu heimildirnar’, in Frændafundur, vol. II, ed.
T. Sigurðardóttir and M. Snædal (Tórshavn, 1997), pp.41–58.
———, ‘Í Uppistovubeitinum: Site and Settlement’, Fróðskaparrit, 45 (1997), 27–44.
Arnórsdóttir, A. S., and H. Þorláksson, ‘Saga heimilis á miðöldum’, in Íslenska söguþingið 28.–31.
maí 1997: Ráðstefnurit, 2 vols (Reykjavík, 1998), I, 31–56.
Astill, G., ‘General Survey, 600–1300’, in The Cambridge Urban History of Britain, vol. I:
600–1540, ed. D. Palliser (Cambridge, 2000), pp.27–49.
———, ‘Towns and Town Hierarchies in Saxon England’, Oxford Journal of Archaeology, 10
(1991), 95–117.
Astill, G., and W. Davies, A Breton Landscape (London, 1997).
Page 341
Ordernr. 061297 P1
326 Bibliography
Aston, M., ‘The Towns of Somerset’, in Anglo-Saxon Towns in Southern England, ed. J. Haslam
(Chichester, 1984), pp.167–201.
Aubrey, J., Monumenta Brittanica ([n.d.]).
Baker, N., and R . Holt, ‘The City of Worcester in the Tenth Century’, in St Oswald of Worcester:
Life and Influence, ed. N. Brooks and C. Cubitt (Leicester, 1996), pp.129–46.
Bange, F., ‘L’ager et la villa: structures du paysage et du peuplement dans la région Mâconnaise à
la fin du haut moyen âge (IX e– XIe siècles)’, Annales ESC, 41 (1984), 529–69.
Barbero, A., and M. Vigil, La formación del feudalismo en la Península Ibérica (Barcelona, 1978).
Barrios García, A., and I. Martín Viso, ‘Reflexiones sobre el poblamiento rural altomedieval en el
Norte de la Península Ibérica’, Studia Historica: Historia Medieval, 18–19 (2000–01), 53–83.
Barrow, G., ‘Frontier and Settlement: Which Influenced Which? England and Scotland,
1100–1300’, in Medieval Frontier Societies, ed. R . Bartlett and A. Mackay (Oxford, 1989),
pp.3–21.
Barrow, J., ‘Churches, Education and Literacy in Towns, 600–1300’, in The Cambridge Urban
History of Britain, vol. I: 600–1540, ed. D. Palliser (Cambridge, 2000), pp.127–52.
Barthélemy, D., ‘Comment 1’, Past and Present, 152 (1996), 196–205.
———, La mutation de l’an mil, a-t-elle eu lieu? Servage et chevalerie dans la France des X e et XIe
siècles (Paris, 1997).
———, ‘La mutation féodale, a-t-elle eu lieu? (Note critique)’, Annales ESC, 47 (1992), 767–77.
———, ‘Note sur les cartulaires de Marmoutier (Touraine) au XIe siècle’, in Les Cartulaires, ed.
O. Guyotjeannin, L. Morelle and L. Parisse (Paris, 1992), pp.247–59.
Bartlett, R ., The Hanged Man: A Story of Miracle, Memory and Colonialism in the Middle Ages
(Princeton, 2004).
Bartlett, R ., and A. Mackay, eds, Medieval Frontier Societies (Oxford, 1989).
Barton, R . E., Lordship in the County of Maine, c. 890–1160 (Woodbridge, 2004).
Bassett, S., ‘The Administrative Landscape of the Diocese of Worcester in the Tenth Century’,
in St Oswald of Worcester: Life and Influence, ed. N. Brooks and C. Cubitt (Leicester, 1996),
pp.147–73.
———, ‘Anglo-Saxon Birmingham’, Midland History, 25 (2000), 1–27.
———, Anglo-Saxon Coventry and its Churches, Dugdale Society Occasional Papers, 41 (2001).
———, ‘Continuity and Fission in the Anglo-Saxon Landscape: The Origins of the Rodings
(Essex)’, Landscape History, 19 (1997), 25–42.
———, The Origins of the Parishes of the Deerhurst Area (Deerhurst Lecture 1997, 1998).
———, The Wootton Wawen Project: Interim Report No. 1 (University of Birmingham, School
of History, 1983).
of History, 1984).
of History, 1985).
of History, 1986).
of History, 1987).
of History, 1988).
Page 342
Ordernr. 061297 P1
Bibliography 327
of History, 1989).
of History, 1991).
Bately, J. M., ‘The Compilation of the Anglo-Saxon Chronicle, 60 BC to AD 890: Vocabulary as
Evidence’, Proceedings of the British Academy, 64 (1978), 93–127.
Beltrán Lloris, F., ‘Un espejismo historiográfico: las organizaciones gentilicias hispanas’, in I
Congreso Peninsular de Historia Antigua (Santiago de Compostela, 1986), ed. G. Pereira
Menaut, 3 vols (Santiago de Compostela, 1988), II, 197–237.
———, ‘La hospitalidad celtibérica: una aproximación desde la epigrafía latina’, Palaeohispanica,
1 (2001), 35–62.
———, ‘Los pactos de hospitalidad de la Hispania Citerior: una valoración histórica’, in La
Península Ibérica hace 2000 años, ed. L. Hernández Guerra, L. Sagredo San Eustaquio and
J. Mª. Solana Sáinz (Valladolid, 2001), pp.393–99.
———, ‘Una variante provincial del hospitium: pactos de hospitalidad y concesión de la ciudad-
anía local en la Hispania Tarraconense’, in Epigrafía y Sociedad en Hispania durante el Alto
Imperio: estructuras y relaciones sociales (Madrid-Alcalá de Henares, 2000), ed. S. Armani,
B. Hurlet-Martineau and A. U. Stylow, Acta Complutensia, 4 (Alcalá, 2003), pp.35–62.
Beltrán Lloris, F., and others, El tercer bronce de Botorrita (Contrebia Belaisca) (Zaragoza, 1996).
Bender, H., ‘Archaeological Perspectives on Rural Settlement in Late Antiquity in the Rhine and
Danube Area’, in Urban Centers and Rural Contexts in Late Antiquity, ed. T. S. Burns and
J. W. Eadie (East Lansing, MI, 2001), pp.147–61.
Biddle, M., ‘Towns’, in The Archaeology of Anglo-Saxon England, ed. D. M. Wilson (London,
1976), pp.91–150.
Biddle, M., and D. Keene, ‘General Survey and Conclusions’, in Winchester in the Early Middle
Ages, ed. M. Biddle (Oxford, 1976), pp.449–508.
Bisson, T. N., ‘The “Feudal Revolution”’, Past and Present, 142 (1994), 6–42.
———, ‘Reply’, Past and Present, 155 (1997), 208–25.
Blackburn, M., ‘“Productive” Sites and the Pattern of Coin Loss in England, 600–1180’, in
Markets in Early Medieval Europe: Trading and ‘Productive’ Sites, 650–850, ed. T. Pestell and
K. Ulmschneider (Macclesfield, 2003), pp.20–36.
Blair, J., ‘Anglo-Saxon Bicester: The Minster and the Town’, Oxoniensia, 67 (2002), 133–40.
———, Anglo-Saxon Oxfordshire (Stroud, 1994).
———, The Church in Anglo-Saxon Society (Oxford, 2005).
———, ‘Ecclesiastical Organization and Pastoral Care in Anglo-Saxon England’, Early Medieval
Europe, 4 (1995), 193–212.
———, ‘Introduction: From Minster to Parish Church’, in Minsters and Parish Churches: The
Local Church in Transition, 950–1200, ed. J. Blair, Oxford University Committee for
Archaeology, monograph 17 (Oxford, 1988), pp.1–19.
———, ‘Small Towns, 600–1270’, in The Cambridge Urban History of Britain, vol. I: 600–1540,
ed. D. Palliser (Cambridge, 2000), pp.245–70.
Blaising, J.-M., ‘Yutz (57), archéologie d’un terroir des âges de metaux aux XIX e siècle’,
Archaeologia Mosellana, 4 (2002), 185–217.
Boesch Gajano, S., ‘Reliques et pouvoir’, in Les Reliques: Objects, Cultes, Symboles, ed. E. Bozóky
and A. M. Helvétius (Turnhout, 1999), pp.255–69.
Page 343
Ordernr. 061297 P1
328 Bibliography
Bohigas Roldán, R ., ‘Fuentes arqueológicas y organización social del espacio en el reino de

Castilla’, in Del Cantábrico al Duero: Trece Estudios sobre Organización Social del Espacio en
los siglos VIII a XIII, ed. J. A. García de Cortázar (Santander, 1999), pp.75–121.
———, Yacimientos Arqueológicos Medievales del Sector Central de la Montaña Cantábrica
(Santander, 1986).
Bohigas Roldán, R ., and others, ‘Carta Arqueológica de la Provincia de Burgos: Partidos Judiciales
de Sedano y Villarcayo’, Kobie, 14 (1984), 7–91.
———, ‘Evolución de las formas y funciones de la arquitectura en el Norte de Hispania: el caso
de Tedeja’, in 3º Congresso de Arqueología Peninsular, 8 vols (Oporto, 2001), VI, 555–68.
———, ‘Tedeja y el control político del territorio del Norte Burgalés en Época Tardorromana,
Visigoda, Alto y Plenomedieval’, in V Congreso de Arqueología Medieval Española, 2 vols
(Valladolid, 2001), I, 49–56.
Böhner, K., Die fränkischen Altertümer des Triererlandes, 2 vols, Germanische Dennkmäler der
Völkerwanderungszeit, Serie B, Band 1 (Berlin, 1958).
Bonney, D., ‘Early Boundaries in Wessex’, in Archaeology and the Lansdcape, ed. P. J. Fowler
(London, 1972), pp.168–86.
Boon, G. C., ‘The Latest Objects from Silchester, Hants.’, Medieval Archaeology, 3 (1959), 79–88.
Bourdieu, P., ‘The Kabyle House or the World Reversed’, in Bourdieu, Algeria 1960 (Cambridge,
1979), pp.133–53.
Bowen, H. C., The Archaeology of Bokerley Dyke, ed. B. N. Eagles (London, 1990).
Bradley, R ., ‘The South Oxfordshire Grim’s Ditch and its Significance’, Oxoniensia, 33 (1968),
1–13.
Bragason, Ú. ‘The Word and the World: A Resonant Textual Fragment of Íslendinga Saga’, paper
given at Sagas and Societies, International Conference at Borgarnes, Iceland 5–9 September
2002, <http://w210.ub.uni-tuebingen.de/dbt/volltexte/2004/1081/pdf/28_ulf~1.pdf>
Brewer, M. B., ‘On the Social Origins of Human Nature’, in The Message of Social Psychology, ed.
C. McGarty and S. A. Haslam (Oxford, 1997), pp.54–62.
Britnell, R. H., The Commercialisation of English Society, 1000–1500, 2nd edn (Manchester, 1996).
Brogiolo, G., ‘Towns, Forts and the Countryside: Archaeological Models for Northern Italy in
the Early Lombard Period’, in Towns and their Territories between Late Antiquity and the
Early Middle Ages, ed. G. P. Brogiolo, N. Christie and N. Gauthier (Leiden, 2000),
pp.299–323.
Brooke, C. N. L., ‘Rural Ecclesiastical Institutions in England: The Search for their Origins’, Setti-
mane di studio del centro italiano di studi sull’ alto medioevo, 28 (1982), 685–711.
Brooks, N. P., ‘The Administrative Background to the Burghal Hidage’, in The Defence of Wessex:
The Burghal Hidage and Anglo-Saxon Fortifications, ed. D. Hill and A. R . Rumble (Man-
chester, 1996), pp.128–50.
———, ‘The Development of Military Obligations in Eighth- and Ninth-Century England’, in
England before the Conquest, ed. P. Clemoes and K. Hughes (Cambridge, 1971), pp.69–84.
———, ‘England in the Ninth Century: The Crucible of Defeat’, Transactions of the Royal
Historical Society, 5th series, 29 (1979), 1–20.
Brown, D. H., ‘Pottery and Late Saxon Southampton’, Proceedings of the Hampshire Field Club
and Archaeological Society, 50 (1995), 127–52.
———, Pottery in Medieval Southampton c.1066–1510 (York, 2002).
Burillo Mozota, F., Los Celtíberos: Etnias y estados (Barcelona, 1998).
Page 344
Ordernr. 061297 P1
Bibliography 329
Callander Murray, A. C., Germanic Kinship Structure: Studies in Law and Society in Antiquity and
the Early Middle Ages (Toronto, 1983).
Callow, C. P., ‘Landscape, Tradition and Power in a Region of Medieval Iceland: Dalir
c.900–c.1262’ (unpublished doctoral thesis, University of Birmingham, 2001).
Cambridge, E., and D. Rollason, ‘The Pastoral Organization of the Anglo-Saxon Church: A
Review of the “Minster Hypothesis”’, Early Medieval Europe, 4 (1995), 87–104.
Cammarosano, P., Italia medievale: Geografia e storia delle fonti scritte (Rome, 1991).
Campbell, J., The Anglo-Saxon State (London, 2000).
———, ‘The Late Anglo-Saxon State: A Maximum View’, Proceedings of the British Academy,
87 (1994), 39–65.
———, ‘Power and Authority’, in The Cambridge Urban History of Britain, vol. I: 600–1540,
ed. D. Palliser (Cambridge, 2000), pp.51–78.
Campillo Cueva, J., ‘Los despoblados medievales de la Honor de Sedano (Burgos)’, Kobie, 24
(1997), 125–68.
Cantarella, G. M., ‘Lo spazio dei monaci’, Settimane di studio del centro italiano di studi sull’alto
medioevo (Uomo e spazio nell’alto medioevo), 50 (2003), 804–47.
Carlstein, T., Time Resources, Society and Ecology (London, 1982).
Carmichael, D. L., ‘Places of Power: Mescalero Apache Sacred Sites and Sensitive Areas’, in Sacred
Sites, Sacred Places, ed. D. Carmichael, J. Hubert, B. Reeves and A. Schanche (London, 1994),
pp.89–98.
Carroll, M., Romans, Celts and Germans: The German Provinces of Rome (Stroud, 2001).
Casey, J., The Kingdom of Valencia in the Seventeenth Century (Cambridge, 1979).
Cassidy-Welch, M., Monastic Spaces and their Meanings: Thirteenth-Century English Cistercian
Monasteries (Turnhout, 2001).
Castellanos, S., and I. Martín Viso, ‘The Local Articulation of Central Power in the North of the
Iberian Peninsula (500–1000)’, Early Medieval Europe, 13 (2005), 1–42.
Cepas Palanca, A., Crisis y continuidad en la Hispania del siglo III, Anejos de Archivo Español de
Arqueología, 17 (Madrid, 1997).
Chapelot, J., and R . Fossier, The Village and House in the Middle Ages (London, 1980).
Chisholm, M., Rural Settlement and Land Use, 3rd edn (London, 1979).
Christaller, W., ‘How I Discovered the Theory of Central Places: A Report about the Origin of
Central Places’, in Man, Space, and Environment: Concepts in Contemporary Human
Geography, ed. P. W. English and R . C. Mayfield (London, 1972), pp.601–10.
Clark, A. J., ‘The Nature of Wansdyke’, Antiquity, 32 (1958), 89–97.
Clutton-Brock, T. H., and P. H. Harvey, ‘Primate Ecology and Social Organization’, in Readings
in Sociobiology, ed. T. H. Clutton-Brock and P. H. Harvey (Reading and San Francisco,
1978), pp.342–83.
Colte Hoare, R ., The Ancient History of Wiltshire, 2 vols (London, 1812–21).
Constable, G., ‘The Ceremonies and Symbolism of Entering Religious Life and Taking the
Monastic Habit, from the Fourth to the Twelfth Century’, Settimane di studio del centro itali-
ano di studi sull’alto medioevo (Segni e Riti della Chiesa Altomedievale), 33 (1987), 771–834.
Cooper, W., Wootton Wawen: Its History and Records (Leeds, 1936).
Corbier, M., ‘Grande proprietà fondaria e piccole azienda: La Gallia settentrionale in epoca
Romana’, in Società Romana e Impero Tardoantico, vol. III: Le merci, gli insediamenti, ed.
A. Giardina (Rome, 1986), pp.687–702, with notes at pp.897–903.
Page 345
Ordernr. 061297 P1
330 Bibliography
Corradini, R ., M. Diesenberger and H. Reimitz, eds, The Construction of Communities in the

Early Middle Ages: Texts, Resources and Artefacts (Leiden, 2003).
Cosgrove, D., and S. Daniels, eds, The Iconography of Landscape (Cambridge, 1988).
Courcelle, P., ‘La vision cosmique de saint Benoît’, Revue des études augustiniennes, 13 (1967),
97–117.
Crawford, O. G. S., ‘The Anglo-Saxon Bounds of Bedwyn and Burbage’, Wiltshire Archaeological
and Natural History Magazine, 131 (1942), 280–301.
———, Archaeology in the Field (London, 1953).
Cubitt, C., ‘The Tenth-Century Benedictine Reform’, Early Medieval Europe, 6 (1997), 77–94.
Cüppers, H., ed., Trier: Kaiserrezidenz und Bishofssitz: Die Stadt in spätantiker und frühchrist-
licher Zeit (Mainz, 1984).
Curchin, L. A., ‘Vici and Pagi in Roman Spain’, Revue d’Etudes Anciennes, 87 (1985), 327–43.
Darby, H., Domesday England (Cambridge, 1977).
Davies, W., ‘The Early Middle Ages and Spanish Identity’, in Power and Identity: Essays Presented
to Rees Davies, ed. H. Pryce and J. Watts (Oxford, forthcoming).
———, ‘“Protected Space” in Britain and Ireland in the Middle Ages’, in Scotland in Dark Age
Britain, ed. B. E. Crawford (St Andrews, 1996), pp.1–19.
———, Small Worlds: The Village Community in Early Medieval Brittany (London, 1988).
Davies, W., and P. Fouracre, eds, The Settlement of Disputes in Early Medieval Europe (Cam-
bridge, 1986).
de Moxó, S., Repoblación y Sociedad en la España cristiana medieval (Madrid, 1979).
de Nie, G., ‘Seeing and Believing in the Early Middle Ages: A Preliminary Investigation’, in
de Nie, Word, Image and Experience: Dynamics of Miracle and Self-Perception in Sixth-
Century Gaul (Aldershot, 2003), ch. VII.
de Palol, P., Clunia: Historia de la ciudad y guía de las excavaciones, 6th edn (Burgos, 1994).
———, ‘Noves dades arqueologiques sobre els darrers segles de Clunia’, in de Palol and others,
Clunia 0: Studia Varia Cluniensia (Burgos, 1991), pp.295–300.
———, ‘Un vidrio tallado, con temas cristianos, de Clunia’, in de Palol and others, Clunia 0:
Studia Varia Cluniensia (Burgos, 1991), pp.347–54.
Dickens, P., Society and Nature: Changing our Environment, Changing Ourselves (Cambridge,
2004).
———, Society and Nature: Towards a Green Social Theory (New York and London, 1992).
Díez Herrera, C., La formación de la sociedad feudal en Cantabria (Santander, 1990).
Dimier, A., ‘Le mot locus dans le sense de monastère’, Revue Mabillon, 62 (1972), 133–54.
Dolley, R ., and D. M. Metcalf, ‘The Reform of the English Coinage under Edgar’, in Anglo-Saxon
Coins, ed. R . Dolley (London, 1961), pp.136–68.
Domínguez Bolaños, A., and J. Nuño González, ‘Reflexiones sobre sistemas defensivos
tardoantiguos en la Cuenca del Duero: a propósito de la muralla de “El Cristo de San
Esteban”, Muelas del Pan (Zamora)’, in La Hispania de Teodosio, ed. R . Teja and C. Pérez, 2
vols (Salamanca, 1997), II, 435–50.
D’Ors, A., Epigrafía Jurídica de la España romana (Madrid, 1953).
———, ‘Un nuevo dato para la historia de la llamada Termancia’, in Estudios dedicados a
Menéndez Pidal, 3 vols (Madrid, 1951), II, 567–81.
Douglas, M., Purity and Danger: An Analysis of the Concepts of Pollution and Taboo, new edn
(London, 2002).
Page 346
Ordernr. 061297 P1
Bibliography 331
Drew, C. D., ‘The Manors of the Iwerne Valley, Dorset’, Proceedings of the Dorset Natural History
and Antiquarian Field Club, 69 (1948), 45–50.
Dugdale, W., Antiquities of Warwickshire, 2nd edn in 2 vols (London, 1730).
———, Monasticon Anglicanum, ed. J. Caley, H. Ellis and B. Bandinel, 6 vols in 8 (London,
1849).
Durand, M., Archéologie du cimetière médiévale au sud-ouest de l’Oise: relations avec l’habitat et
évolution des rites et des pratiques funéraires du VIèm e au XVIèm e siècle, Revue Archéologique de
Picardie, numéro spéciale (Amiens, 1988).
———, ‘Cimetières et habitats: l’émergence du village au sud-ouest du département de l’Oise’,
Revue Archéologique de Picardie, 3–4 (1988), 153–60.
Durliat, J., Les finances publiques de Dioclétien aux Carolingiens (284–890), Beihefte der Francia,
21 (Sigmaringen, 1990).
Dutton, P. E., The Politics of Dreaming in the Carolingian Empire (Lincoln and London, 1994).
Eagles, B. N., ‘Anglo-Saxon Presence and Culture in Wiltshire’, in Roman Wiltshire and After:
Papers in Honour of Ken Annable, ed. P. Ellis (Devizes, 2001), pp.199–233.
———, ‘Evidence for Settlement in the Fifth to Seventh Centuries AD ’, in The Medieval
Landscape of Wessex, ed. M. Aston and C. Lewis, Oxbow Monograph, 46 (Oxford, 1994),
pp.13–32.
Edson, E., Mapping Time and Space: How Medieval Mapmakers Viewed their World (London,
1997).
Einarsson, Á., O. Hansson and O. Vésteinsson, ‘An Extensive System of Medieval Earthworks in
Northeast Iceland’, Archaeologia Islandica, 2 (2002), 61–73.
Elton, H., Warfare in Roman Europe, 350–425 (Oxford, 1996).
Engelbert, P., Die Vita Sturmi des Eigil von Fulda: Literarkritisch-historische Untersuchung und
Edition (Marburg, 1968).
Escalona Monge, J., ‘Acerca de la territorialidad en la Castilla altomedieval: Tres casos
significativos’, in Historia social, pensamiento historiográfico y Edad Media, ed. I. Loring García
(Madrid, 1997), pp.217–44.
———, ‘Comunidades, territorios y poder condal en la Castilla del Duero en el siglo X ’, Studia
Historica: Historia Medieval, 18–19 (2000–01), 85–120.
———, ‘Family Memories: Inventing Alfonso I of Asturias’, in Building Legitimacy: Political
Discourses and Forms of Legitimation in Medieval Societies, ed. I. Alfonso, H. Kennedy and
J. Escalona (Leiden, 2004), pp.223–62.
———, ‘Jerarquización social y organización del espacio: Bosques y pastizales en la Sierra de
Burgos’, in Los rebaños de Gerión: Pastores y trashumancia en Iberia antigua y medieval, ed.
J. Gómez-Pantoja, Collection de la Casa de Velázquez, 73 (Madrid, 2001), pp.109–37.
———, ‘Poblamiento y organización territorial en el sector oriental de la cuenca del Duero en
la Alta Edad Media’, in III Congreso de Arqueología Medieval Española, 2 vols (Oviedo, 1992),
II, 448–55.
———, ‘De “señores y campesinos” a “poderes feudales y comunidades”: Elementos para definir
la articulación entre territorio y clases sociales en la Alta Edad Media castellana’, in Comuni-
dades locales y poderes feudales en la Edad Media, coord. I. Álvarez Borge (Logroño, 2001),
pp.117–55.
———, Sociedad y territorio en la Alta Edad Media Castellana: La Formación del Alfoz de Lara,
British Archaeological Reports, International Series, 1079 (Oxford, 2002).
Page 347
Ordernr. 061297 P1
332 Bibliography
Espinosa Ruiz, U., Epigrafía romana de la Rioja (Logroño, 1986).

Estepa Díez, C., ‘El alfoz castellano en los siglos IX al XII’, in En la España Medieval, IV: Estudios
Dedicados al Profesor D. Ángel Ferrari Núñez, 2 vols (Madrid, 1984), II, 305–41.
———, ‘Burgos en el contexto del nacimiento de la ciudad medieval castellano-leonesa’, in La
ciudad de Burgos: Congreso de Historia de Burgos (Madrid, 1985), pp.23–55.
Everitt, A., Continuity and Colonization: The Evolution of Kentish Settlement (Leicester, 1986).
Fabech, C., and J. Ringtved, eds, Settlement and Landscape: Proceedings of a Conference in Århus,
Denmark May 4–7 1998 (Moesgård, 1999).
Fairbrother, J., Faccombe Netherton: Excavations of a Saxon and Medieval Manorial Complex
(London, 1990).
Faith, R ., The English Peasantry and the Growth of Lordship (Leicester, 1997).
Farmer, S., Communities of St Martin: Legend and Ritual in Medieval Tours (Ithaca and London,
1991).
Fernández Mier, M., Génesis del Territorio en la Edad Media: Arqueología del Paisaje y Evolución
Histórica en la Montaña Asturiana (Oviedo, 1999).
Fernie, E., The Architecture of the Anglo-Saxons (London, 1983).
Filloy Nieva, I., ‘Yacimiento de Las Ermitas (Espejo)’, Arkeoikuska (1995), 295–302.
Finley, M. I., The Ancient Economy, 2nd edn (London, 1985).
Fleming, R ., ‘Lords and Labour’, in From the Vikings to the Normans: The Short Oxford History
of the British Isles (800–1100), ed. W. Davies (Oxford, 2003), pp.107–37.
———, ‘Monastic Lands and England’s Defence in the Viking Age’, English Historical Review,
100 (1985), 247–65.
———, ‘Rural Elites and Urban Communities in Late-Saxon England’, Past and Present, 141
(1993), 3–37.
Fletcher, R ., The Conversion of Europe: From Paganism to Christianity 371–1386 AD (London,
1997).
Fontaine, J., ‘Isidore de Séville et la mutation de l’encyclopédisme antique’, in Tradition et actua-
lité chez Isidore de Séville, ed. J. Fontaine, special issue, Cahiers d’histoire mondiale / Journal of
world history / Cuadernos del Historia Mundial, 9.3 (1966), 519–38.
Fossier, R ., ‘Rural Economy and Country Life’, in The New Cambridge Medieval History, vol. III:
c.900–1024, ed. T. Reuter (Cambridge, 1999), pp.27–63.
Foucault, M., Discipline and Punish: The Birth of the Prison (Harmondsworth, 1977).
———, ‘Espace, savoir et pouvoir’, in Foucault, Dits et Écrit 1954–1988, 4 vols (Paris, 1994), IV ,
270–85.
———, ‘The Eye of Power’, in Foucault, Power/Knowledge: Selected Interviews and Other
Writings 1972–1977 (New York, 1980), pp.146–65.
Fouracre, P., The Age of Charles Martel (London, 2000).
Fourquin, G., ‘Les temps de croissance’, in Histoire de la France rurale, vol. I: Des origines à 1340,
ed. G. Duby (Paris, 1975), pp.370–547.
Fowler, P. J., Landscape Plotted and Pieced: Landscape History and Local Archaeology in Fyfield
and Overton, Wiltshire (London, 2000).
———, ‘Wansdyke in the Woods: An Unfinished Roman Military Earthwork for a Non-event’,
in Roman Wiltshire and After: Papers in Honour of Ken Annable, ed. P. Ellis (Devizes, 2001),
pp.179–98.
Fox, A., and C. Fox, ‘Wansdyke Reconsidered’, Archaeological Journal, 115 (1958), 1–48.
Page 348
Ordernr. 061297 P1
Bibliography 333
Gantier, O., ‘Recherches sur les possessions et les prieurés de l’Abbaye de Marmoutier du X e au
e
XIII siècle’, Revue Mabillon, 53 (1963), 93–110.
García Camino, I., Arqueología y Poblamiento en Bizkaia (siglos VI– XII): La Configuración de la
Sociedad Feudal (Bilbao, 2002).
García de Cortázar, J. A., ‘Organización social del espacio burgalés en la Alta Edad Media’, in II
Jornadas Burgalesas de Historia: Burgos en la Alta Edad Media (Burgos, 1991), pp.19–74.
———, ed., Organización social del espacio en la España Medieval: La Corona de Castilla en los
siglos VIII al XV (Barcelona, 1985).
García de Cortázar, J. A., and E. Peña Bocos, ‘De alfoces, aldeas y solares en la Castilla de los siglos
IX al XI: ¿una formalización “feudal” del espacio?’, in Miscellania en Homenatge al P. Agustí
Altisent (Tarragona, 1991), pp.183–202.
García González, J. J., ‘Del cerro al castillo: El cerro de Burgos de la Antigüedad a la Edad Media’,
Cuadernos Burgaleses de Historia Medieval, 2 (1995), 71–166.
García Merino, M. C., Población y poblamiento en la Hispania Romana: El Conventus Cluniensis
(Valladolid, 1975).
Gardiner, M., and C. Greatorex, ‘Archaeological Excavations in Steyning, 1992–95: Further
Evidence for the Evolution of a Late Saxon Small Town’, Sussex Archaeological Collections,
135 (1997), 143–71.
Garnsey, P., and R . Saller, The Roman Empire: Economy, Society, Culture (London, 1987).
Gaultier Dalché, P., ‘De la glose à la contemplation: Place et fonction de la carte dans les manu-
scrits du haut Moyen Âge’, Settimane di studio del centro italiano di studi sull’alto medioevo
(Testo e immagine nell’alto medioevo), 41 (1994), 693–764.
———, ‘Tradition et renouvellement dans la représentation de l’espace géographique au IX e
siècle’, Studi medievali, 3rd series, 24 (1983), 121–65.
Gaunt, D., Femiljeliv i Norden (Stockholm, 1983).
Geary, P. J., ‘Ethnic Identity as a Situational Construct in the Early Middle Ages’, Mitteilungen
der anthropologischen Gesellschaft in Wien, 113 (1983), 15–26.
———, Furta Sacra: Thefts of Relics in the Central Middle Ages, rev. edn with new preface
(Princeton, 1991).
———, The Myth of Nations: The Medieval Origins of Europe (Princeton, 2002).
———, Phantoms of Remembrance: Memory and Oblivion at the End of the First Millennium
(Princeton, 1995).
Geertz, C., ‘Two Types of Ecosystems’, in Man, Space and Environment: Concepts in Contempo-
rary Human Geography, ed. P. W. English and R . C. Mayfield (London, 1972), pp.165–80.
Gelling, M., ‘The Place-Name Volumes for Worcestershire and Warwickshire: A New Look’, in
Field and Forest: An Historical Geography of Warwickshire and Worcestershire, ed. T. R . Slater
and P. J. Jarvis (Norwich, 1982), pp.59–78.
Gelling, M., and A. Cole, The Landscape of Place-Names (Stamford, 2000).
Genicot, L., Rural Communities in the Medieval West (Baltimore, 1990).
Gerhold, W., Armut und Armenfürsorge im mittelalterlichen Island, Skandinavistische Arbeiten,
18 (Heidelberg, 2002).
Giardina, A., ‘Roma e il Caucaso’, Settimane di studio del centro italiano di studi sull’alto medioev
(Il Caucaso: cerniera fra culture dal Mediterraneo alla Persia (secoli IV – XI)), 43 (1996),
85–141.
Page 349
Ordernr. 061297 P1
334 Bibliography
Giddens, A., Runaway World: How Globalisation is Reshaping our Lives, 2nd edn (London,
2002).
Gilchrist, R ., Gender and Material Culture: The Archaeology of Religious Women (London and
New York, 1994).
Gimeno, H., and M. Mayer, ‘Una propuesta de identificación epigráfica: Lara de los Infantes /
Nova Augusta’, Chiron, 23 (1993), 313–21.
Gimeno García-Lomas, R ., ‘El conjunto de cerámicas medievales de Santa María de Hito’, in I
Congreso de Arqueología Medieval Española, 5 vols (Zaragoza, 1985), II, 385–401.
Godman, P., and R . Collins, eds, Charlemagne’s Heir: New Perspectives on the Reign of Louis the
Pious (814–840) (Oxford, 1990).
Goetz, H.-W., ‘Serfdom and the Beginnings of a “Seigneurial System” in the Carolingian Period:
A Review of the Evidence’, Early Medieval Europe, 2 (1993), 29–51.
Goffart, W., Barbarians and Romans, 418–584: The Techniques of Accommodation (Princeton,
1980)
———, Caput and Colonate: Towards a History of Late Roman Taxation (Toronto, 1974).
———, ‘From Roman Taxation to Medieval Seigneurie: Three Notes’, Speculum, 47 (1972),
15–87, 373–94.
———, ‘Old and New in Merovingian Taxation’, Past and Present, 96 (1982), 3–21.
Gómez Fraile, J. Mª., ‘Celtiberia en las fuentes grecolatinas: Replanteamiento conceptual de un
paradigma obsoleto’, Polis, 8 (1996), 143–206.
Gómez-Pantoja, J., ‘Historia de dos ciudades: Capera y Clunia’, in Économie et territoire en
Lusitanie romaine, ed. J.-G. Gorges and F. Rodríguez Martín, Collection de la Casa de
Velázquez, 65 (Madrid, 1999), pp.91–108.
———, ‘Pastio agrestis: Pastoralismo en Hispania romana’, in Los rebaños de Gerión: Pastores y
trashumancia en Iberia antigua y medieval, ed. J. Gómez-Pantoja, Collection de la Casa de
Velázquez, 73 (Madrid, 2001), pp.177–213.
González, Mª. C., and J. Santos, ‘La epigrafía del conventus Cluniense, I: Las estelas vadinienses’,
Memorias de Historia Antigua, 6 (1984), 85–111.
González Rodríguez, M. C., Las unidades organizativas indígenas en el área indoeuropea de la
Península Ibérica (Vitoria, 1986).
Gover, J. E. B., A. Mawer and F. M. Stenton, The Place-Names of Warwickshire, English Place-
Name Society, 13 (Cambridge, 1936).
Gover, J. E. B., and others, The Place Names of Wiltshire (Cambridge, 1939)
Graham, A., and S. Davies, Excavations in Trowbridge, Wiltshire, 1977 and 1986–88 (Salisbury,
1993).
Graham, D., ‘The Foundation Charter of the Chapel of St John the Evangelist at Henley-in-
Arden, 1367’, Warwickshire History, 12 (2003–04), 154–68.
Granger, G.-G., La pensée de l’espace (Paris, 1999).
Green, D., C. Haselgrove and M. Spriggs, eds, Social Organization and Settlement: Contributions
from Anthropology, Archaeology and Geography, 2 vols, British Archaeological Reports,
International Series, 47 (Oxford, 1978).
Green, H. S., ‘Wansdyke: Excavations 1966 to 1970’, Wiltshire Archaeological and Natural
History Magazine, 67 (1971), 129–46.
Grenville, J., Medieval Housing (Leicester, 1997).
Page 350
Ordernr. 061297 P1
Bibliography 335
Guerreau, A., L’avenir d’un passé incertain: Quelle histoire du Moyen Âge au XXIe siècle? (Paris,
2001).
———, ‘Le champ sémantique de l’espace dans la vita de saint Maieul (Cluny, début du XIe
siècle)’, Journal des Savants (1997), 363–420.
———, ‘Chasse’, in Dictionnaire Raisonné de l’Occident Médiéval, ed. J. Le Goff and J.-C.
Schmitt (Paris, 1999), pp.166–78.
———, ‘Structure et évolution des représentations de l’espace dans le haut Moyen Âge occiden-
tal’, Settimane di studio del centro italiano di studi sull’alto medioevo (Uomo e spazio nell’alto
medioevo), 50 (2003), 91–115.
Guillaume, J., ‘Les nécropoles mérovingiennes de Dieue/Meuse (France)’, Acta Praehistorica et
Archaeologica, 5–6 (1974–75), 211–349.
Gutiérrez González, J. A., Fortificaciones y Feudalismo en el Origen y Formación del Reino Leonés
(siglos IX – XIII) (Valladolid, 1995).
———, ed., Peñaferruz (Gijón): El Castillo de Curiel y su Territorio (Gijón, 2003).
———, ‘Sobre los orígenes de la sociedad asturleonesa: aportaciones desde la arqueología del
territorio’, Studia Historica: Historia Medieval, 16 (1998), 173–97.
Haas, W., ‘Foris apostolus – intus monachus: Ansgar als Mönch und “Apostel des Nordens”’,
Journal of Medieval History, 11 (1985), 1–30.
Haggett, P., A. D. Cliff and A. Frey, Locational Models (London, 1977; = vol. 1 of 2nd edn of
P. Haggett, Locational Analysis in Human Geography, 2 vols).
Haley, E., Migration and Economy in Roman Spain (Barcelona, 1991).
Hall, T. A., Minster Churches in the Dorset Landscape, British Archaeological Reports, British
Series, 304 (Oxford, 2000).
Halldórsson, B., ‘Atli’, in Rit Björns Halldórssonar í Sauðlauksdal (Reykjavík, 1983), pp.49–213.
Halsall, G., ‘Archaeology and the Late Roman Frontier in Northern Gaul: The So-called Födera-
tengräber Reconsidered’, in Grenze und Differenz im früheren Mittelalter, ed. W. Pohl and
H. Reimitz (Vienna, 2000), pp.167–80.
———, Barbarian Migrations and the Roman West (Cambridge, forthcoming)
———, ‘Burial, Ritual and Merovingian Society’, in The Community, the Family and the Saint:
Patterns of Power in Early Medieval Europe, ed. J. Hill and M. Swan (Turnhout, 1998),
pp.325–38.
———, ‘Burial Writes: Graves, “Texts” and Time in Early Merovingian Northern Gaul’, in
Erinnerungskultur im Bestattungsritual: Archäologisch-Historisches Forum, ed. J. Jarnut
(Darmstadt, 2004), pp.61–74.
———, ‘The Origins of the Reihengräberzivilisation: Forty Years On’, in Fifth-Century Gaul:
A Crisis of Identity?, ed. J. F. Drinkwater and H. Elton (Cambridge, 1992), pp.196–207.
———, Settlement and Social Organization: The Merovingian Region of Metz (Cambridge, 1995).
———, Warfare and Society in the Barbarian West, 450–900 (London, 2003).
Hamerow, H., Early Medieval Settlements: The Archaeology of Rural Communities in North-West
Europe 400–900 (Oxford, 2002).
Harley, J. B., ‘Maps, Knowledge, and Power’, in The Iconography of Landscape, ed. D. Cosgrove
and S. Daniels (Cambridge, 1988), pp.277–312.
Harrison, D., Medieval Space: The Extent of Microspatial Knowledge in Western Europe during the
Middle Ages (Lund, 1996).
Page 351
Ordernr. 061297 P1
336 Bibliography
Hase, P. H., ‘The Church in the Wessex Heartlands’, in The Medieval Landscape of Wessex, ed.
M. Aston and C. Lewis, Oxbow Monograph, 46 (Oxford, 1994), pp.47–81.
———, ‘The Mother Churches of Hampshire’, in Minsters and Parish Churches: The Local
Church in Transition, 950–1200, ed. J. Blair, Oxford University Committee for Archaeology,
monograph 17 (Oxford, 1988), pp.45–66.
Haselgrove, C., and C. Scull, ‘Early Medieval Rural Settlement in North-East France: Field
Survey Evidence from the Aisne Valley, Picardy’, in Early Medieval Europe 1992: A
Conference on Medieval Archaeology in Europe, 21 st–24 th Sept 1992 at the University of York:
Pre-printed Papers, 8 vols (York, 1992), VIII, 207–12.
Haslam, J., ‘The Towns of Wiltshire’, in Anglo-Saxon Towns in Southern England, ed. J. Haslam
(Chichester, 1984), pp.87–147.
Heene, K., ‘Gender and Mobility in the Low Countries: Travelling Women in Thirteenth-
century Exempla and Saints’ Lives’, in The Texture of Society: Medieval Women in the
Southern Low Countries, ed. E. E. Kittell and M. A. Suydam (London, 2004), pp.31–49.
Hennebicque, R ., ‘Espaces sauvages et chasses royales dans le Nord de la France, VIIe– IX e siècle’,
Revue du Nord, 62 (1980), 35–60.
Hernández Guerra, L., and L. Sagredo San Antonio, La Romanización del territorio de la actual
provincia de Palencia (Valladolid, 1998).
Heyse, E., Hrabanus Maurus’ Enzyklopädie ‘De rerum naturis’ (Munich, 1969).
Hill, D., ‘Athelstan’s Urban Reforms’, Anglo-Saxon Studies in Archaeology and History, 11 (2000),
173–85.
———, An Atlas of Anglo-Saxon England (Oxford, 1981).
———, ‘The Burghal Hidage: The Establishment of a Text’, Medieval Archaeology, 13 (1969),
84–92.
———, ‘The Calculation and the Purpose of the Burghal Hidage’, in The Defence of Wessex
Wessex: The Burghal Hidage and Anglo-Saxon Fortifications, ed. D. Hill and A. R . Rumble
(Manchester, 1996), pp.92–97.
———, ‘The Construction of Offa’s Dyke’, Antiquaries Journal, 65 (1985), 140–42.
Hill, D., and A. R . Rumble, eds, The Defence of Wessex: The Burghal Hidage and Anglo-Saxon
Fortifications (Manchester, 1996).
Hill, D., and M. Worthington, Offa’s Dyke, History and Guide (Stroud, 2003).
Hinchliffe, J., ‘Excavations at Grim’s Ditch, Mongewell, 1974’, Oxoniensia, 40 (1975), 122–35.
Hindle, B. P., Roads, Tracks and their Interpretation (London, 1993).
Hinton, D., ‘Coins and Commercial Centres in Anglo-Saxon England’, in Anglo-Saxon Monetary
History, ed. M. Blackburn (Leicester, 1986), pp.11–26.
———, ‘The Large Towns, 600–1300’, in The Cambridge Urban History of Britain, vol. I:
Hinton, D. A., and C. J. Webster, ‘Excavations at the Church of St Martin, Wareham, 1985–86,
and “Minsters” in South-east Dorset’, Proceedings of the Dorset Natural History and Archaeo-
logical Society, 109 (1987), 47–54.
Hlawitschka, E., Franken, Alemannen, Bayern und Burgunder in Oberitalien (774–962): Zum
Verständnis der fränkischen Königsherrschaft in Italien (Freiburg im Breisgau, 1960).
Hodder, I., The Present Past (London, 1982).
Hodges, R ., The Anglo-Saxon Achievement (London, 1989).
Page 352
Bibliography 337
Hoffmann, D., Das Spätrömische Bewegungsheer und die Notitia Dignitatum, 2 vols, Epigra-
phische Studien, 7.1 and 7.2 (Düsseldorf, 1969–70).
Hooke, D., The Landscape of Anglo-Saxon England (Leicester, 1998).
———, Warwickshire Anglo-Saxon Charter-Bounds (Woodbridge, 1999).
Hopkins, K., Conquerors and Slaves (Cambridge, 1978).
Hoskins, W. G., and H. P. R . Finberg, Devonshire Studies (London, 1952).
Hostetter, E., and T. N. Howe, The Romano-British Villa at Castle Copse, Great Bedwyn
(Bloomington, 1997).
Hreiðarsdóttir, E., and O. Vésteinsson, Fornleifaskráning á Húsavík [FS166–01171] (Fornleifa-
stofnun Íslands, 2002).
Hughes, M., ‘Grimsditch and Cuthwulf’s Expedition to the Chilterns in AD 571’, Antiquity, 5
(1931), 291–314.
Imízcoz, J. M., ‘Comunidad de valle y feudalismo en el norte de la península: Algunas preguntas
desde el Valle de Baztán’, in Señorío y feudalismo en la Península Ibérica (siglos XII– XIX ), ed.
E. Sarasa Sánchez and E. Serrano Martín, 4 vols (Zaragoza, 1993), III, 69–86.
Innes, M., State and Society in the Early Middle Ages: The Middle Rhine Valley, 400–1000
(Cambridge, 2000).
Jakobsson, S., ‘Griðamál á ófriðaröld’, in Íslenska söguþingið 28.–31. maí 1997: Ráðstefnurit, 2 vols
(Reykjavík, 1998), I, 117–34.
Jarrett, J., ‘Power Over Past and Future: Abbess Emma and the Nunnery of Sant Joan de les
Abadesses’, Early Medieval Europe, 12 (2003), 229–56.
Jimeno, A., Epigrafía romana de Soria (Soria, 1980).
Joffroy, R ., Le Cimetiére Mérovingien de Lavoye (Paris, 1974).
Jones, G. R . J., ‘The Multiple Estate as a Model Framework for Tracing Early Stages in the
Evolution of Rural Settlement’, in L’Habitat et les paysages ruraux d’Europe, ed. F. Dussart
(Liège, 1971), pp.251–67.
———, ‘Multiple Estates and Early Settlement’, in Medieval Settlement: Continuity and Change,
ed. P. H. Sawyer (London, 1976), pp.15–40.
Jones, R . F. J., ‘A False Start? The Roman Urbanization of Western Europe’, World Archaeology,
19 (1987), 47–57.
Júlíusdóttir, S., Den kapabla familjen i det isländska samhället: en studie om lojalitet,
äktenskapsdynamik och psykosocial anpassning (Göteborg, 1993).
Kain, R . J. P., and R . R . Oliver, The Historic Parishes of England and Wales (Colchester, 2001).
Kastner, J., Historiae fundationum monasteriorum: Frühformen monasticher Institutionsgeschichts-
schreibung im Mittelalter, Münchener Beiträge zur Mediävistik und Renaissance-Forschung,
18 (Munich, 1978).
Keene, D., ‘Towns and the Growth of Trade’, in The New Cambridge Medieval History, vol. IV :
c.1024–1198, Part I, ed. D. Luscombe and J. Riley-Smith (Cambridge, 2004), pp.47–85.
Kristjánsson, L., Íslenzkir sjávarhættir, vol. II (Reykjavík, 1982).
Kulturhistorisk leksikon for nordisk middelalder, 22 vols (Reykjavík, 1956–78).
Lagazzi, L., Segni sulla terra: Determinazione dei confini e percezione dello spazio nell’alto medioevo
(Bologna, 1991).
Langdon, J., Horses, Oxen and Technological Innovation: The Use of Draught Animals in English
Farming 1066–1500 (Cambridge, 1986).
Lapidge, M., ‘The Archetype of Beowulf’, Anglo-Saxon England, 29 (2000), 5–41.
Page 353
Ordernr. 061297 P1
338 Bibliography
Lárusdóttir, B., and E. Hreiðarsdóttir, Fornleifaskráning í Reykjahreppi, 2 vols (Fornleifastofnun

Íslands, 2002–03).
Lárusson, Ó., Byggð og saga (Reykjavík, 1944).
Lavinson, S. C., ‘Space: Linguistic Expression’, in International Encyclopedia of the Social and
Behavioral Sciences, 29 vols (Amsterdam, 2001), XXII, 14749–52.
Le Goff, J., ‘Discorso di chiusura’, Settimane di studio del centro italiano di studi sull’alto medioevo
(Popoli e paesi nella cultura altomedievale), 29 (1983), 805–38.
Le Roux, P., L’armée romaine et l’organisation des provinces ibériques d’Auguste à l’invasion de 409
(Paris, 1982).
———, ‘Vicus et castellum en Lusitanie sous l’Empire’, Studia Historica: Historia Antigua, 10–11
(1992–93), 151–60.
Leakey, R ., and R . Lewin, Origins Reconsidered: In Search of What Makes Us Human (London,
1992).
Lecanda Esteban, J. Á., ‘Mijangos: la aportación de la epigrafía y el análisis arqueológico al conoci-
miento de la transición a la alta edad media en Castilla’, in Visigodos y Omeyas: Un Debate
entre la Antigüedad Tardía y la Alta Edad Media, ed. L. Caballero and P. Mateos (Madrid,
2000), pp.181–206.
———, ‘Mijangos: arquitectura y ocupación visigoda en el norte de Burgos’, in II Congreso de
Arqueología Peninsular (Zamora, 1996), pp.415–34.
Lee, A. D., ‘The Army’, in The Cambridge Ancient History, vol. XIII: The Late Empire, A .D .
337–425, ed. A. Cameron and P. Garnsey (Cambridge, 1998), pp.211–37.
Lennard, R., Rural England 1086–1135 (Oxford, 1959).
López Paz, P., La Ciudad Romana Ideal, vol. I.1: El territorio (Santiago de Compostela, 1994).
López Quiroga, J., ‘Fluctuaciones del poblamiento y hábitat “fortificado” de Altura en el Noroeste
de la Península Ibérica (ss. V – IX )’, in Mil Anos de Fortificaçôes na Península Ibérica e no
Magreb (500–1500) (Lisbon, 2001), pp.83–91.
Louis, E., ‘A De-Romanised Landscape in Northern Gaul: The Scarpe Valley from the Fourth to
the Ninth Century AD ’, in Recent Research on the Late Antique Countryside, ed. W. Bowden,
L. Lavan and C. Marchado (Leiden, 2004), pp.479–504.
Löwe, H., ‘Westliche Peregrination und Mission: Ihr Zusammenhang mit den Länder und
Völkerurkundlichen Kenntnissen des früheren Mittelalters’, Settimane di studio del centro
italiano di studi sull’alto medioevo (Popoli e paesi nella cultura altomedievale), 29 (1983),
327–72.
Loyn, H., ‘Towns in Late Anglo-Saxon England: The Evidence and Some Possible Lines of
Enquiry’, in England before the Conquest, ed. P. Clemoes and K. Hughes (Cambridge 1971),
pp.115–28.
Lozovsky, N., The Earth is Our Book: Geographical Knowledge in the Latin West ca. 400–1000
(Ann Arbor, 2000).
Lutz, M., ‘Le domaine gallo-romaine de Saint-Ulrich (II)’, Gallia, 30 (1972), 41–82.
MacLean, S., Kingship and Politics in the Late Ninth Century: Charles the Fat and the End of the
Carolingian Empire (Cambridge, 2003).
MacMullen, R ., ‘The Celtic Renaissance’, Historia, 14 (1965), 93–104.
———, ‘The Epigraphic Habit in the Roman Empire’, American Journal of Philology, 103
(1982), 233–46.
Madox, T., Formulare Anglicanum (London, 1702).
Page 354
Ordernr. 061297 P1
Bibliography 339
Maitland, F. W., Domesday Book and Beyond (Cambridge, 1897).

Major, A. F., and E. J. Burrow, The Mystery of Wansdyke (Cheltenham, 1926).
Malmberg, T., Human Territoriality: Survey of Behavioural Territories in Man with Preliminary
Analysis and Discussion of Meaning (The Hague, 1980).
Mañanes, T., and J. M. Solana, Ciudades y vías romanas en la Cuenca del Duero (Valladolid,
1985).
Markus, R . A., ‘How on Earth Could Places Become Holy? Origins of the Christian Idea of Holy
Places’, Journal of Early Christian Studies, 2 (1994), 257–71.
Martín, J. L., Orígenes de la Orden Militar de Santiago (1170–1195) (Barcelona, 1974).
Martín Viso, I., Fragmentos del Leviatán: La articulación política del espacio zamorano en la alta
edad media (Zamora, 2002).
———, ‘Organización episcopal y poder entre la Antigüedad Tardía y el Medioevo (siglos V – XI):
Las diócesis de Calahorra, Oca y Osma’, Iberia, 2 (1999), 151–90.
———, ‘Pervivencia y transformación de los sistemas castrales en la formación del feudalismo en
la Castilla del Ebro’, in Comunidades locales y poderes feudales en la Edad Media, coord.
I. Álvarez Borge (Logroño, 2001), pp.255–88.
———, Poblamiento y estructuras sociales en el norte de la Península Ibérica (siglos VI– XIII)
(Salamanca, 2000).
———, ‘Poblamiento y sociedad en la transición al feudalismo en Castilla: Castros y aldeas en
la Lora burgalesa’, Studia Historica: Historia Medieval, 3 (1995), 3–45.
———, ‘Poder político y estructura social en la Castilla altomedieval: el condado de Lantarón
(siglos VIII– XI)’, in Los Espacios de Poder en la España Medieval, coord. J. I. de la Iglesia Duarte
(Logroño, 2002), pp.533–52.
Martínez Díez, G., Pueblos y alfoces burgaleses de la repoblación (Valladolid, 1987).
Martínez Sopena, P., La Tierra de Campos Occidental: Poblamiento, poder y comunidad del siglo
X al XIII (Valladolid, 1985).
Mawer, A., and F. M. Stenton, The Place-Names of Worcestershire, English Place-Name Society,
4 (Cambridge, 1927).
Mellor, M., Medieval Ceramic Studies in England (London, 1994).
Menéndez Pidal, R ., En torno a la lengua vasca (Buenos Aires, 1962).
Metcalf, D. M., An Atlas of Anglo-Saxon and Norman Coin Finds, c.973–1086 (London, 1998).
———, ‘Monetary Circulation in Southern England in the First Half of the Eighth Century’,
in Sceattas in England and the Continent, ed. D. Hill and D. M. Metcalf (Oxford, 1984),
pp.27–71.
Meyvaert, P., ‘Peter the Deacon and the Tomb of Saint Benedict’, Revue bénédictine, 65 (1955),
3–70.
Millar, F., The Emperor in the Roman World, 31 BC – AD 337 (London, 1977).
———, The Roman Empire and its Neighbours (London, 1967).
Miller, W. I., Bloodtaking and Peacemaking (Chicago, 1990).
Millet, M., The Romanisation of Britain (Cambridge, 1990).
Moore, R . I., ‘Family, Community and Cult on the Eve of the Gregorian Reform’, Transactions
of the Royal Historical Society, 5th series, 30 (1980), 49–69.
———, The First European Revolution c. 970–1215 (Oxford, 2000).
Moreland, J., ‘The Significance of Production in Eighth-Century England’, in The Long Eighth
Century, ed. I. L. Hansen and C. Wickham (Leiden, 2000), pp.69–104.
Page 355
Ordernr. 061297 P1
340 Bibliography
Morrison, S. S., Women Pilgrims in Late Medieval England: Private Piety as Public Performance
(London, 2000).
Mrozek, S., ‘À propos de la répartition chronologique des inscriptions latines dans le Haut-
Empire’, Epigraphica, 35 (1973), 114.
Myres, J. N. L., ‘Wansdyke and the Origin of Wessex’, in Essays in British History, ed. H. R .
Trevor-Roper (London, 1964), pp.1–27.
Näsman, U., ‘Exchange and Politics: The Eighth–Early Ninth Century in Denmark’, in The Long
Eighth Century, ed. I. L. Hansen and C. Wickham (Leiden, 2000), pp.35–68.
———, ‘The Justinianic Era of South Scandinavia: An Archaeological View’, in The Sixth
Century: Production, Distribution and Demand, ed. R . Hodges and W. Bowden (Leiden,
1998), pp.255–78.
Nelson, J., Charles the Bald (London, 1992).
Nightingale, J., Monasteries and Patrons in the Gorze Reform (Oxford, 2001).
Noisette, P., ‘Usages et représentations de l’espace dans la Regula Benedicti: Une nouvelle
approche des significations historiques de la Règle’, Regulae Benedicti Studia: Annuarium
internationale, 14–15 (1985–86), 69–80.
Oexle, O. G., ‘Memoria und Memorialüberlieferung im früheren Mittelalter’, Frühmittelalterliche
Studien, 10 (1976), 70–95.
Óla, Á., Þúsund ára sveitaþorp: Úr sögu Þykkvabæjar í Rangárþingi (Reykjavík, 1962).
Ólason, V., and others, eds., Íslensk Bókmenntasaga, 2 vols (Reykjavík, 1992–93).
O’Loughlin, T., ‘Living in the Ocean’, in Studies in the Cult of Columba, ed. C. Bourke (Dublin,
1997), pp.11–23.
Olsen, M., Farms and Fanes of Ancient Norway (Oslo, 1928).
Oman, C., ‘Wansdyke’, Archaeological Journal, 87 (1930), 60–70.
Ortega Valcárcel, J., La Cantabria Rural: Sobre la Montaña (Santander, 1987).
O’Sullivan, D., ‘Space, Silence and Shortage on Lindisfarne: The Archaeology of Ascetism’, in
Image and Power in the Archaeology of Early Medieval Britain: Essays in Honour of Rosemary
Cramp, ed. H. Hamerow and A. MacGregor (Oxford, 2001), pp.33–52.
Palliser, D. M., ‘The “Minster Hypothesis”: A Case Study’, Early Medieval Europe, 5 (1996),
207–14.
Palmer, B., ‘The Hinterlands of Three Southern English Emporia: Some Common Themes’, in
Markets in Early Medieval Europe: Trading and ‘Productive’ Sites, 650–850, ed. T. Pestell and
K. Ulmschneider (Macclesfield, 2003), pp.48–60.
Passmore, A. D., ‘The Age and Origin of the Wansdyke’, Antiquaries Journal, 4 (1924), 26–29.
Pastor, R ., C. Estepa, J. A. García de Cortázar, J. L. Abellán and J. L. Martín, Sánchez Albornoz
a debate (Valladolid, 1993).
Paul, J., ‘Expression et perception du temps d’après l’enquête sur les miracles de Louis d’Anjou’,
in Temps, mémoire, tradition au Moyen Âge (Aix-en-Provence, 1983), pp.19–41.
Peel, R . F., ‘Local Intermarriage and Stability of Rural Population in the English Midlands’,
Geography, 27 (1942), 22–30.
Peña Bocos, E., ‘Ecclesia y Monasterium, elementos de ordenación de la sociedad de la Castilla
altomedieval’, in Señorío y feudalismo en la Península Ibérica (siglos XII– XIX ), ed. E. Sarasa
Sánchez and E. Serrano Martín, 4 vols (Zaragoza, 1993), III, 379–98.
Penco, G., ‘Un elemento della mentalità monastica medievale: la concezione dello spazio’,
Benedictina, 35 (1988), 35–71.
Page 356
Ordernr. 061297 P1
Bibliography 341
Pereira Menaut, G., ‘Che cos’è un munus?’, Athenaeum, 1 (2004), 169–215.

———, ‘Cognatio Magilancum: Una forma de organización indígena de la Hispania
indoeuropea’, in Actas del V Coloquio sobre Lenguas y Culturas Prerromanas de la Península
Ibérica, ed. J. Untermann and F. Villar (Salamanca, 1993), pp.411–24.
———, ‘Formación técnica vs. humanismo: Aproximación crítica’, Revista Mientras Tanto,
68–69 (1997), 135–48.
Pereira Menaut, G., and J. Santos Yanguas, ‘Ensayo de sistematización de la epigrafía romana de
Asturias: Las unidades gentilicias’, Boletín del Instituto de Estudios Asturianos, 105–06 (1982),
87–110.
———, ‘Sobre la romanización del noroeste de la península Ibérica: las inscripciones con
mención de origo personal’, Revista de Guimaraes (1980), 120–22.
Perring, D., Town and Country in England: Frameworks for Archaeological Research (York, 2002).
Pflaum, H. G., Les carrières procuratoriennes équestres sous le Haut Empire romain, 3 vols (Paris,
1961).
Pitt Rivers, A. H. L. F., Excavations in Bokerly and Wansdyke Dorset and Wilts, 1888–1891
(printed privately, 1892).
Pohl, W., ‘The Construction of Communities and the Persistence of Paradox: An Introduction’,
in The Construction of Communities in the Early Middle Ages: Texts, Resources and Artefacts,
ed. R . Corradini, M. Diesenberger and H. Reimitz (Leiden, 2003), pp.1–15.
———, ‘Telling the Difference: Signs of Ethnic Identity’, in Strategies of Distinction: The
Construction of Ethnic Communities, 300–800, ed. W. Pohl and H. Reimitz (Leiden, 1998),
pp.17–69.
Pohl, W., and H. Reimitz, eds, Strategies of Distinction: The Construction of Ethnic Communities,
300–800 (Leiden, 1998).
Pollard, J., and A. J. Reynolds, Avebury: The Biography of a Landscape (Stroud, 2002).
Porlan, A., Los nombres de Europa (Madrid, 1998).
Pratesi, A., ‘Ubi corpus beati Clementis papae et martyris requiescit’, in Contributi per una storia
dell’Abruzzo adriatico nel medioevo (Chieti, 1992), pp.115–31.
Rainbird, P., ‘Oldaport and the Anglo-Saxon Defence of Devon’, Proceedings of the Devon
Archaeological Society, 56 (1988), 153–64.
Ramírez Sánchez, M. E., Epigrafía y organización social en la region celtibérica: los grupos de
parentesco (Las Palmas de Gran Canaria, 2001).
Remensnyder, A., Remembering Kings Past: Monastic Foundation Legends in Medieval Southern
France (Ithaca and London, 1995).
———, ‘Topographies of Memory: Center and Periphery in High Medieval France’, in Concepts
of the Past: Ritual, Memory, Historiography, ed. G. Althoff, J. Fried and P. J. Geary (Washing-
ton and Cambridge, 2002), pp.193–214.
Remotti, F., P. Scarduelli and U. Fabietti, eds, Centri, ritualità, potere: Significati antropologici
dello spazio (Bologna, 1989).
Renfrew, C., ed., The Explanation of Culture Change: Models in Prehistory (London, 1973).
Renn, D., Norman Castles in Britain, 2nd edn (London, 1973).
Reuter, T., ‘Comment 3’, Past and Present, 155 (1997), 177–95.
———, ‘The Making of England and Germany, 850–1050: Points of Comparison and Differ-
ence’, in Medieval Europeans, ed. A. Smyth (Basingstoke, 1998), pp.53–70.
Page 357
Ordernr. 061297 P1
342 Bibliography
Rey Losada, A., La Ciudad Romana Ideal, vol. II: El grupo poblacional (Santiago de Compostela,
2003).
Reyes, F., Población y sociedad en el Valle del Duero: Duratón y Riaza en la alta Edad Media, siglos
VI al XI (Madrid, 2002).
Reyes, F., and M. L. Menéndez Robles, ‘Aspectos ideológicos en la despoblación del Valle del
Duero’, in Historiografía de la Arqueología y de la Historia Antigua en España (siglos XVIII– XX )
(Madrid, 1991), pp.203–07.
Reynolds, A. J., ‘Boundaries and Settlements in Later Sixth to Eleventh-Century England’, in
Boundaries in Early Medieval Britain, ed. D. Griffiths, A. J. Reynolds and S. Semple (Oxford,
2003), pp.98–136.
———, ‘From Pagus to Parish: Territory and Settlement in the Avebury Region from the Late
Roman Period to the Domesday Survey’, in The Avebury Landscape: Aspects of the Field
Archaeology of the Marlborough Downs, ed. G. Brown, D. Field and D. McOmish (Oxford,
2005), pp.164–80.
———, Later Anglo-Saxon England (Stroud, 1999).
Ripoll, G., and J. Arce, ‘The Transformation and End of Roman villae in the West
(Fourth–Seventh Centuries): Problems and Perspectives’, in Towns and their Territories
between Late Antiquity and the Early Middle Ages, ed. G. P. Brogiolo, N. Christie and
N. Gauthier (Leiden, 2000), pp.63–114.
Robinson, P., ‘The Treasure Act’, Trilithon, 46 (2002), 3.
Rollason, D., ‘Monasteries and Society in Early Medieval Northumbria’, in Monasteries and
Society in Medieval Britain, ed. B. Thompson (Stamford, 1999), pp.59–74.
Rosenwein, B. H., To Be the Neighbor of Saint Peter: The Social Meaning of Cluny’s Property
909–1049 (Ithaca and London, 1989).
Rosser, G., ‘The Anglo-Saxon Gilds’, in Minsters and Parish Churches: The Local Church in
Transition, 950–1200, ed. J. Blair, Oxford University Committee for Archaeology, mono-
graph 17 (Oxford, 1988), pp.31–34.
Rumble, A. R ., ‘Diplomatic Sub Sections’, in The Defence of Wessex: The Burghal Hidage and
Anglo-Saxon Fortifications, ed. D. Hill and A. R . Rumble (Manchester, 1996), pp.69–74.
Russo, D. G., Town Origins and Development in Early England, c. 400–950 A .D . (Westport, CN,
1998).
Sack, R . D., Human Territoriality: Its Theory and History (Cambridge, 1986).
Sáenz de Urturi, F., ‘Castros de Lastra: XX campaña de excavaciones’, Arkeoikuska (1993), 53–62.
Salinas de Frías, M., ‘Sobre las formas de propiedad comunal de la cuenca del Duero en época
prerromana’, Veleia, 6 (1989), 103–10.
Salway, P., Roman Britain (Oxford, 1981).
Sánchez-Albornoz, C., Despoblación y Repoblación del Valle del Duero (Buenos Aires, 1966).
Santos Yanguas, J., ‘1985–1994: Un decenio fructífero en la investigación de las estructuras
sociales indígenas del área indoeuropea’, Veleia, 12 (1995), 125–49.
Sawyer, P. H., Anglo-Saxon Charters: An Annotated Hand List and Bibliography (London, 1968).
———, ‘The Royal Tun in Pre-Conquest England’, in Ideal and Reality in Frankish and Anglo-
Saxon Society, ed. P. Wormald (Oxford, 1983), pp.273–99.
Semple, S., ‘Burials and Political Boundaries in the Avebury Region, North Wiltshire’, in Boun-
daries in Early Medieval Britain, ed. D. Griffiths, A. J. Reynolds and S. Semple (Oxford,
2003), pp.72–91.
Page 358
Ordernr. 061297 P1
Bibliography 343
———, ‘A Fear of the Past: The Place of the Prehistoric Burial Mound in the Ideology of the
Middle and Later Anglo-Saxon England’, World Archaeology, 30 (1998), 109–26.
Sennis, A., ‘Tradizione monastica e racconto delle origini in Italia centrale (secoli XI– XII)’,
Mélanges de l’École française de Rome, 115 (2003), 181–211.
Il sepolcro di S. Benedetto, Miscellanea cassinese, 27 (Montecassino, 1951).
Shorters, I. A., A Wasting Historical Asset? A Comparative Study of Grave Memorials at Wootton
Wawen, King’s Norton and Birmingham, c.1700–1940, British Archaeological Reports,
British Series, 366 (Oxford, 2004).
Sigurgeirsson, M. Á., O. Vésteinsson, R . Edvardsson and T. Horsley, Fornleifarannsókn á
Steinboga í Mývatnssveit 2002 [FS189–02072] (Fornleifastofnun Íslands, 2003).
Sigurðsson, J. V., Frá Goðorðum til ríkja (Reykjavík, 1989).
———, ‘Heimili, þingmennska, frændur og vernd á þjóðveldisöld’, in Íslenska söguþingið 28.–31.
maí 1997: Ráðstefnurit, 2 vols (Reykjavík, 1998), I, 107–16.
Simmer, A., Le cimetière Mérovingien d’Audun-le-Tiche: archéologie d’aujourd’hui no. 1, Mémoire
de l’Association française d’Archéologie Mérovingienne, 2 (Paris, 1988).
———, La nécropole d’Ennery (Moselle) (Woippy, 1992).
Slater, T., ‘The South-West of England’, in The Cambridge Urban History of Britain, vol. I:
Slater, T. R ., ‘More on the Wolds’, Journal of Historical Geography, 5 (1979), 213–18.
Smith, A., ‘Medieval Archaeological Features at Church Street, Milborne Port’, Somerset
Archaeology and Natural History, 146 (2003), 37–46.
Smith, J. Z., Map Is Not Territory (Leiden, 1978).
Sonnlechner, C., ‘The Establishment of New Units of Production in Carolingian Times: Making
Early Medieval Sources Relevant for Environmental History’, Viator, 35 (2004), 21–48.
Southern, R . W., The Making of the Middle Ages (London, 1967).
Spiegel, G. M., The Past as Text: The Theory and Practice of Medieval Historiography (Baltimore,
1997).
Spratt, D. A., Linear Earthworks of the Tabular Hills, Northeast Yorkshire (Sheffield, 1989).
Stafford, P., ‘The “Farm of One Night” and the Organization of King Edward’s Estates in
Domesday’, Economic History Review, 33 (1980), 491–502.
———, Unification and Conquest (London, 1989).
Stein, F., ‘Die Bevölkerung des Saar-Mosel-Raumes am Übergang von der Antike zum
Mittelalter: Überlegungen zum Kontinuitätsproblem aus archäologischer Sicht’, Archaeologia
Mosellana, 1 (1989), 89–195.
———, ‘Les tombes d’un chef franc et de sa famille à Güdingen: Considérations sur le rôle de
l’aristocratie dans l’implantation franque entre la Meuse et la Sarre’, Saarbrücker Studien und
Materialen zur Altertumskunde, 1 (1992), 117–44.
Stenton, F. M., Anglo-Saxon England, 3rd edn (Oxford, 1971).
———, The Early History of the Abbey of Abingdon (Reading, 1913).
———, ‘The Road System of Medieval England’, Economic History Review, 7 (1936–37), 1–21.
Stevenson, W. H., ‘Trinoda Necessitas’, English Historical Review, 29 (1914), 689–702.
Stukeley, W., Itinerarium Curiosum, Centuria, vol. II (London, 1776).
Svorou, S., The Grammar of Space (Amsterdam, 1994).
Talmy, L., ‘How Language Structures Space’, in Spatial Orientation: Theory, Research, and
Application, ed. H. L. Pick, Jr, and L. P. Acredolo (New York, 1983), pp.225–82.
Page 359
Ordernr. 061297 P1
344 Bibliography
Taylor, C. S., ‘The Date of Wansdyke’, Transactions of the Bristol and Gloucestershire Archaeo-
logical Society, 27 (1908), 131–55.
Taylor, H. M., and J. Taylor, Anglo-Saxon Architecture, 2 vols (Cambridge, 1965).
Thoroddsen, Þ., Lýsing Íslands, 4 vols (Reykjavík, 1908–22).
Thorsteinsson, A., ‘On the Development of Faroese Settlements’, in Proceedings of the Eighth
Viking Congress Århus 24–31 August 1977, ed. H. Bekker-Nielsen, P. Foote and O. Olsen
(Odense, 1981), pp.189–202.
Tomlin, R . S. O., ‘The Army of the Late Empire’, in The Roman World, ed. J. Wacher (London,
1987), pp.107–23.
Tönnies, F., Community and Civil Society, trans. J. Harris and M. Hollis (Cambridge, 2001; first
published 1887 as Gemeinschaft und Gesellschaft).
Toulmin Smith, L., ed., The Itinerary of John Leland in or about the Years 1535–43, 5 vols
(Carbondale, 1964).
Ulmschneider, K., ‘Settlement, Economy and the “Productive” Site in Anglo-Saxon Lincolnshire,
AD 650–780’, Medieval Archaeology, 44 (2000), 53–79.
Unión Académica Internacional (Comité Español), Hoja K-30: Madrid, Caesaraugusta-Clunia,
Tabula Imperii Romana (Madrid, 1993).
Unwin, T., ‘Rural Marketing in Medieval Nottinghamshire’, Journal of Historical Geography, 7
(1981), 231–51.
Unzueta, M., and A. Martínez, ‘Proyecto de variante y túnel entre las Conchas de Haro y el Cruce
de Briñas’, in Arqueología de Urgencia en Álava, 1989–1993 (Vitoria, 1994), pp.43–60.
Uríbarri, J. L., J. M. Martínez González and I. Leis, Primeros asentamientos humanos en la ciudad
de Burgos (Burgos, 1987).
van Ossel, P., Établissements ruraux de l’antiquité tardive dans le nord de la Gaule, 51 e Supplément
à Gallia (Paris, 1992).
———, ‘Insécurité et militarisation en Gaule du Nord au Bas-Empire: L’exemple des campagnes’,
Revue du Nord, 77 (1995), 27–36.
Vázquez, R ., ‘Castros, castillos y torres en la organización social del espacio en Castilla: el espacio
del Arlanza al Duero (siglos IX a XIII)’, in Del Cantábrico al Duero: Trece Estudios sobre
Organización Social del Espacio en los siglos VIII a XIII, ed. J. A. García de Cortázar (Santander,
1999), pp.351–73.
VCH, Warwickshire, vol. III, ed. P. Styles (London, 1945).
Verhulst, A., The Rise of Cities in Northwest Europe (Cambridge, 1999).
Vésteinsson, O., The Christianization of Iceland: Priests, Power, and Social Change 1000–1300
(Oxford, 2000).
———, ‘The Formative Phase of the Icelandic Church c. 990–1240 AD ’, in Church Centres, ed.
H. Þorláksson (forthcoming).
———, Forn kirkja og grafreitur á Neðra-Ási í Hjaltadal (Reykjavík, 2000).
———, ‘Íslenska sóknaskipulagið og samband heimila á miðöldum’, in Íslenska söguþingið
28.–31. maí 1997: Ráðstefnurit, 2 vols (Reykjavík, 1998), I, 147–66.
———, ed., Landscapes of Settlement 2002 [FS218–02261] (Fornleifastofnun Íslands, 2004)
———, ‘Patterns of Settlement in Iceland: A Study in Prehistory’, Saga-Book of the Viking
Society, 25 (1998), 1–29.
Vésteinsson, O., T. H. McGovern and C. Keller, ‘Enduring Impacts: Social and Environmental
Aspects of Viking Age Settlement in Iceland and Greenland’, Archaeologia Islandica, 2 (2002),
98–136.
Page 360
Ordernr. 061297 P1
Bibliography 345
Vince, A., Saxon London: An Archaeological Investigation (London, 1990).

Voinot, J., ‘Les fouilles de Chaouilley: Cimetière mérovingien’, Mémoires de la Société
Archéologique Lorraine, 54 (1904), 5–80.
von den Brincken, A. D., ‘Mappa Mundi und Chronographia’, Deutsches Archiv für Erforschung
des Mittelalters, 24 (1968), 118–86.
von Thünen, J. H., Der isolierte Staat in Beziehung auf Landwirtschaft und Nationalökonomie
(Berlin, 1876; first part published Hamburg, 1826), translated in Von Thünen’s Isolated State,
ed. P. Hall (Oxford, 1966).
Ward, B., ‘The Desert Myth: Reflections on the Desert Ideal in Early Cistercian Monasticism’,
in One Yet Two: Monastic Tradition in East and West, Orthodox-Cistercian Symposium
(Kalamazoo, 1976), pp.183–99.
Wenskus, R ., Stammesbildung und Verfassung: Das Werden der frühmittelalterlichen Gentes
(Cologne, 1961).
Werner, J., ‘Fernhandel und Naturalwirtschaft im östlichen Merowingerreich nach archäolog-
ischen und numismatischen Zeugnissen’, Bericht der Römisch-Germanischen Kommission, 42
(1961), 307–46.
White, S. D., ‘Comment 2’, Past and Present, 152 (1996), 205–23.
———, Custom, Kinship, and Gifts to Saints: The ‘Laudatio Parentum’ in Western France
1050–1150 (Chapel Hill, 1988).
———, ‘Feuding and Peace-Making in the Touraine around the Year 1100’, Traditio, 42
(1986), 195–263.
———, ‘Proposing the Ordeal and Avoiding It: Strategy and Power in Western French Litiga-
tion, 1050–1110’, in Cultures of Power: Lordship, Status and Process in Twelfth-Century
Europe, ed. T. Bisson (Philadelphia, 1995), pp.89–123.
———, ‘Tenth-Century Courts at Mâcon and the Perils of Structuralist History: Re-reading
Burgundian Judicial Institutions’, in Conflict in Medieval Europe: Changing Perspectives on
Medieval Culture, ed. W. C. Brown and P. Górecki (Aldershot, 2003), pp.37–68.
Whitelock, D. M., The Audience of Beowulf (Oxford, 1951).
Whittaker, C. R ., and P. Garnsey, ‘Rural Life in the Late Roman Empire’, in The Cambridge
Ancient History, vol. XIII: The Late Empire, A .D . 337–425, ed. A. Cameron and P. Garnsey
(Cambridge, 1998), pp.277–311.
Wickham, C., ‘Comment 4’, Past and Present, 155 (1997), 196–208.
———, ‘European Forests in the Early Middle Ages: Landscape and Land Clearance’, in
Wickham, Land and Power: Studies in Italian and European Social History, 400–1200
(London, 1994), pp.155–99.
———, Framing the Early Middle Ages (Oxford, 2005).
———, ‘Overview: Production, Distribution and Demand, II’, in The Long Eighth Century, ed.
I. L. Hansen and C. Wickham (Leiden, 2000), pp.345–77.
———, ‘Problems of Comparing Rural Societies in Early Medieval Western Europe’, Transac-
tions of the Royal Historical Society, 6th series, 2 (1992), 221–46.
Wightman, E. M., Gallia Belgica (London, 1985).
Wood, I., ‘Christians and Pagans in Ninth-Century Scandinavia’, in The Christianization of
Scandinavia, ed. B. Sawyer, P. Sawyer and I. Wood (Alingsås, 1987), pp.36–67.
———, ‘Conclusion: Strategies of Distinction’, in Strategies of Distinction: The Construction of
Ethnic Communities, 300–800, ed. W. Pohl and H. Reimitz (Leiden, 1998), pp.297–303.
Page 361
Ordernr. 061297 P1
346 Bibliography
———, The Missionary Life: Saints and the Evangelisation of Europe 400–1050 (London, 2001).
———, ‘Teutsind, Witlaic and the History of Merovingian Precaria’, in Property and Power in
the Early Middle Ages, ed. W. Davies and P. Fouracre (Cambridge, 1995), pp.31–52.
Wormald, P., ‘Giving God and King their Due: Conflict and its Regulation in the Early English
State’, in Wormald, Legal Culture in the Early Medieval West (London, 1999), pp.333–57.
———, The Making of English Law: King Alfred to the Twelfth Century (Oxford, 1999).
Wynne-Edwards, V. C., Animal Dispersion in Relation to Social Behaviour (Edinburgh, 1962).
———, ‘Self-regulation in Populations of Red Grouse’, in Malthus Past and Present, ed.
J. Dupâquier, A. Fauve-Chamoux and E. Grebenik (London and New York, 1983),
pp.379–91.
Yeager, J., and M. A. Canuto, ‘Introducing an Archaeology of Communities’, in The Archaeology
of Communities: A New World Perspective, ed. M. A. Canuto and J. Yeager (London and New
York, 2000), pp.1–15.
Yorke, B. A. E., ‘Fact or Fiction? The Written Evidence of the 5th and 6th Centuries AD ’, in
Anglo-Saxon Studies in Archaeology and History, vol. VI (Oxford, 1993), 45–50.
———, ‘The Origins of Mercia’, in Mercia: An Anglo-Saxon Kingdom in Europe, ed. M. P.
Brown and C. A. Farr (London, 2001), pp.13–22.
———, Wessex in the Early Middle Ages (Leicester, 1995).
Zadora-Rio, E., ‘The Making of Churchyards and Parish Territories in the Early Medieval
Landscape of France and England in the 7 th –12 th Centuries: A Reconsideration’, Medieval
Archaeology, 47 (2003), 1–19.
Page 362
Ordernr. 061297 P1
INDEX
Places listed are in England unless otherwise stated.
Abingdon (Oxfordshire) 36, 37 alfoz (see also district, suburbium, terri-

Abonae (Sea Mills, Avon) 42 tory) 148–9, 155–6, 159, 161, 163–6,
Adalhard, abbott 279 171, 173, 181, 183
Adam de Cloptone 17 Alfred, King of Wessex 18, 34, 39, 40,
Adomnán 288 235
advowson 128 All Cannings (Wiltshire) 18
Aðaldælahreppur (Iceland) 51 Allington (Wiltshire) 32
Aðaldalur (Iceland) 46, 48, 51, 59, 62, Allo 290 (n.32)
63, 104 Alne, River (Warwickshire) 124, 127,
agriculture 140
estates 189 Alton Priors (Wiltshire) 32
intensification 234 Alþing 80, 85
stock raising 184 Ancilla/ae 255, 265, 266
Aguilar (Spain) 182 ‘Anglian Collection’ 34
Ainardus 290 (n.32) Anglo-Saxon Chronicle 31, 32, 44
air photography 116 Anjou (France), comital house of 261
Alamanns 290 Anskar, archbishop 279 (n.8), 280,
Al-Andalus (Spain) 145 284, 286 (n.20)
Alans 190 Antonine Wall (Scotland) 42
Alcuin, abbott 278 (n.7) Aquae Sulis (Avon) 17
Aldhelm, bishop 18 Arab 178 (n.33), 179
Aldermaston (Berkshire) 23 conquest 143
Alexander the Great 288 rule in Iberia 146
Alfania (Spain) 170, 173, 174, 181 Arbejal (Spain) 183 (n.54)
Alfonso I, King of Asturias (Spain) 145
Page 363
Ordernr. 061297 P1
348 Index
archaeology 26, 88–91, 145, 148 (n.12), ‘auto-dedition’ 263, 264, 270
148–9, 153–4, 157, 161–2, 165 autonomy 172
excavation 135, 142 Avebury (Wiltshire) 32, 35, 41, 241,
fieldwork 142 242
Arrow, River (Warwickshire) 127, 140 Avon, River
architecture 149, 153 (Bristol, Avon) 15, 16, 17, 35
pre-Romanesque 154 (Warwickshire) 140
Romanesque 153–4 Avonmouth (Avon) 42
Arevaci 195 Auxerre (France) 285
aristocracy 172, 174, 175, 177, 178, Álftanes (Iceland) 101, 105
179, 180, 182, 184, 235, 246 Ámót (Iceland) 111–12
Arlanzón (Spain) Árnesþing (Iceland) 99
River 157, 161 Ásar (Iceland) 99
village 157
Arnarbæli (Iceland) 100 Bachicabo (Spain) 171
Arnarnes (Iceland) 72, 75 Bakkafjörður (Iceland) 99
Arnarstapi (Iceland) 96 Bakki (Iceland) 70, 72, 77, 81, 83
Arnþrúðarsons 73 Bampton (England) 239, 241
Arras (France) 281 baptism 131
assemblies (see Alþing, Leið, várþing) 108 Barnstaple (Devon) 241, 242
Asser 18 Barrio de Valdegovía (Spain) 171
Asterio, Bishop of Oca (Spain) 176 Basse-Yutz (Moselle, France) 215
Asturias (Spain) 188, 191, 196 Bath (Avon) 15, 17, 35, 36, 39
kingdom 144 Bathampton (Avon)
kings 146–7 Camp 17
kingship 145 Down 17
royal chronicles (ninth century) Bathford (Avon) 17
144–5 Bár (Iceland) 111
uplands 144 Báscones de Ebro (Spain) 181
Atapuerca, sierra (Spain) 155 Bearley (Warwickshire) 128–9, 130
Athelstan, King of England 138 Beaudesert (Warwickshire) 124, 133–7
Audun-le-Tiche (Moselle, France), Beauvais (France) 281
Merovingian cemetery 224 Bede, the Venerable 31, 44
Auðbrekka (Iceland) 77, 81, 84 Bedwyn (Wiltshire)
Auðsholt (Iceland) 100 dykes 14, 20, 21, 24, 25
Aufidianus 281 Great 20, 21, 25, 241, 251
Ausín (Spain) 151, 154–5, 157–8, Little 20, 21
163, 165 Belgae 35
alfoz 149, 155, 162, 164 Belmore Copse (Wiltshire) 20
fortress 162 Benedict of Nursia 283, 291
Page 364
Ordernr. 061297 P1
Index 349
Bensington (Oxfordshire) 37 Bricia (Spain) 173

Beorhtwulf, King of Mercia 126, 139 Britons 32
Beowulf 34, 40 Brittany 6–9
Berbeia (Spain) 170, 171, 182 (n.51) Brown’s Barn (Wiltshire) 23
Berkshire 35, 36, 37, 42, 44 Buradón (Spain) 177, 178, 180
Berzosa (Spain) 174 Burbage (Wiltshire) 21
Bicester (Oxfordshire) 239, 241 Bure, River (Oxfordshire) 239
Bilbao (Spain) 191 Burghal Hidage 39, 40, 41, 236, 240,
Bilbilis (Spain) 204 243
Bilibio (Spain) 178, 180 Burgos (Spain) 149, 161, 168
bishopric, see episcopal see district 155–7, 161, 164–5
bishops 118, 122, 125, 126, 131, 132, Burhgred, King of Mercia 37
135, 261 burhs
Bishops Cannings Down (Wiltshire) 18 as forts, military centres 39, 44, 235,
Bjarneyjar (Iceland) 96 236, 240, 243, 244, 246
Björn Gestsson 79 occupation in 245, 246, 248
Björn Hallsson 79 as political centres 236
Björn Ólafsson 74 types of 242, 245, 249, 250, 254
Björn Steinmóðarson 82, 83, 86 burial 124, 131
Blois (France), comital house of 261 fees 126, 127, 128, 129
Bokerley Dyke (Dorset/Wiltshire) 22 rights 126, 300
Boniface, archbishop 275 Bustarbrekka (Iceland) 104
Bordesley Abbey (Worcestershire) bygð 91
128, 129 byggð 102
Borgarfjörður (Iceland) 99, 110 Bægisá (including Ytri- and Syðri-
Borgarhús (Iceland) 59 Bægisá) 77, 81
borough 122, 124, 183 Bænhús 100
boundaries 26, 34, 36, 44, 48, 51, 53, Bær 110
60, 61, 64, 91, 102, 111, 116, 117,
125, 134, 140 Cabeza de San Vicente (Spain) 206
district 158, 160–5 Caesaraugusta (Spain) 198
fields 53, 55 Calleva Atrebatum (Hampshire) 42
markers 10, 296–7, 307 Calne (Wiltshire) 241, 251, 252
village 151–2, 157 Canningas 35
Brandr (Örnólfsson) 74, 77 Cannings Marsh (Wiltshire) 32
Breiðafjörður (Iceland) 96 Cantabria (Spain) 168, 172
Breiðavík (Iceland) 97 mountains 143, 188
Brekka (in Aðaldalur, Iceland) 105 sea 144
Brekka (in Ólafsfjörður, Iceland) 104 Capera (Spain) 199
Brekknakot (Iceland) 52, 59 Caranca (Spain) 182
Page 365
Ordernr. 061297 P1
350 Index
Cardeñuela (Spain) 156–8 Caucasus 288

Carolingians 284 Cebrecos (Spain) 164
expansion in north-east Iberia 144 Celtiberia (Spain) 195
cartularies 120, 125, 139, 256, 262, 276 celtic survival 188, 206
cash 267, 271 cemeteries 105, 106, 111, 124, 125, 126,
Castañares (Spain) 157 127, 128, 129, 131, 173, 175, 176
castella 167, 195, 205–6 census 92, 93
Castella (see Castile, Spain) 168 Central Mountains (Spain) 143, 147
Castella Vetula (Spain) 168, 170 central places 10, 11, 12, 181, 182, 183,
Castello Dei (probably Medina de 303–4
Pomar, Spain) 170 Central Place Theory 89
castellum 255 Cenwulf, King of Mercia 40
Castile (Spain) 143–5, 147–8, 151, 159, Chaouilley (Meurthe-et-Moselle,
160–1, 166, 168, 172, 173 (n.19), France), Merovingian cemetery 221
174, 176, 178, 179, 180, 183, 184 chapel
county of 193 manorial 118, 119, 126, 127, 128,
early medieval 143 129, 132, 135
south-eastern 149 parochial 118, 119, 125, 129, 130,
Castilian (Spain) 132, 135, 137
annals 161 chapelry 132
central lands 147 Charlemagne, King of the Franks 290
central settlements 159 Charles the Bald, King of the Franks
counts, comital dynasty 145, 159 40
county 160 charter bounds 14, 18, 26, 32, 259
local communities 148 of donation 255
parish network 149 Máldagar (Icelandic ecclesiastical
rulers 159, 163 charters) 77, 85
settlement 146 of manumission 266, 267
territorial structures 166 charters 21, 24, 26, 31, 39, 40, 102,
Castillejo (Spain) 206 109, 119, 120, 125, 126, 139, 147,
Castillería (Spain) 183 149, 151, 153, 161, 255, 256, 259,
Castillo, El (Spain) 206 260, 266
castle 180, 182, 183, 184 (n.58) Cheddar (Somerset) 241, 242
Castra 167, 220 Chew, River (Avon) 14, 15
Castrillo de Butrón (Spain) 181 Chichester (Sussex) 39, 241, 248, 252,
Castrillo de Valdelomar (Spain) 173 254
Castros de Lastra (Spain) 182 children 265, 267
Castrovido (Spain) 206 Chippenham (Wiltshire) 37
Catalonia (Spain) 143 Chisbury (Wiltshire) 20, 241, 242
cathedral 117, 135 Wood 20
Page 366
Ordernr. 061297 P1
Index 351
Chittoe (Wiltshire) 26 community 14, 41, 115, 116, 142, 172,

Christchurch (Hampshire) 241, 245, 176, 179, 180, 182, 183, 184, 185,
249, 251, 254 265, 267
Chronica Monasterii Casinensis 291 in Castile (Spain) 143
(n.35) extent 5, 7–8, 9, 10, 301, 302, 306
churches 5, 46, 48, 50, 80, 91, 96, in Iceland 45, 58, 63, 67, 81, 88, 89,
106–9, 100, 105–7, 111–12, 146, 102, 104–9, 111–13
173, 175, 176, 177, 180, 246, 249, local 147–8, 157 (n.30), 166
305 of Marmoutier 264, 268
fabric 120, 122, 124, 125, 126, 127, membership and relationship 157,
129, 130, 135 165–6
Cissa, King of Wessex 36 post-Roman redefinition 163 (n.49)
Cistercians 289 supralocal 158
Civitas 35, 170, 189, 190, 196, 197, township 147–8
204, 255 valley communities 146–7, 158
Merovingian 222, 226 village communities 13, 147, 165–6,
Claverdon (Warwickshire) 130–3, 134 237, 238
Clunia (Spain) 160–1, 164–5, 190–3, Cookham (Berkshire) 36
195–7 complexity
district 163 social 147–8, 166
Cluniac spatial 159
monasticism 260 Compton Dando (Avon) 15
reform 260 Conchas de Haro (Spain) 177
Cluniensis 197 Conches (Eure, France) 126, 127, 128,
Coenred, King of Mercia 125 129, 131
coinage Constantine III, emperor 190
circulation of 244, 248, 250 continuity 31, 168
laws 240, 253 conventus (Roman administrative-
sceattas 238, 244 judicial district) 160, 190
Coincidence of time and space 288 conversion (to Christianity) 106, 112
coliberti 268 Conygar Hill (Avon) 16
colonization of deserted spaces 147 Corbie, monastery (France) 278, 279,
Columba 288 286
Columbanus 279, 280 Cornwall 36, 41
Commentary on Scipio’s Dream 280 Coruña del Conde (Spain) 193
Commonwealth period (Iceland) 56 Corvey, monastery (France) 286
commune (Icelandic) 102, 108, 111–12 cottage 52, 53, 55, 59
Communes (modern French) 212 count 174, 178, 181
county 168, 178, 179, 180
courting 87–8
Page 367
Ordernr. 061297 P1
352 Index
courts 244 Devon 36, 41, 242, 254

Covarrubias (abbey, monastery, Spain) Dicuil 285
164 Diego Rodríguez, Count of Castile
Córdoba (Spain), emirs of 144 (Spain) 161
Crediton (Devon) 246 Dieue-sur-Meuse (Meuse, France),
Crewkerne (Somerset) 241, 248 Merovingian cemetery 221
Cricklade (Wiltshire) 241, 245, 249, 251 diocese 119, 129, 246
Cronaca di Novalesa 282 (n.13) geography of 37
Cubillejo de Lara (Spain) 151, 165 Diocesis Hispaniarum (Spain) 190
Cubillo (Spain) Dispenser family 126
territory 150, 152, 164–5 disputes 261, 265, 266, 268, 270
valley 151–2, 155, 165 district (see also alfoz, suburbium, terri-
Cubillo del Campo (Spain) 151 tory) 147–8, 159–61, 163, 165–6,
Cubillo del César (Spain) 151 167, 168, 171, 173, 180, 181, 183
Cunetio (Wiltshire) 17, 21, 42 boundaries 160, 163–4
curacy 129 centre 159–60
Curtis 217 free population 159
Cynewulf, King of Wessex 37 Dobunni 35
Cædwalla, King of Wessex 36, 40 Domesday Book 13, 125, 132, 133,
134, 135, 240, 246, 248, 251
Dalatangi (Iceland) 98 Doña Santos (Spain) 164
dance 88 donations to monasteries 261
Danelaw, boundary of 43, 44 Dorchester (Dorset) 22, 238, 241, 252
defences 249 Dorchester (Oxfordshire) 33, 36
De mensura orbis terrae 285 Dorset 41, 251
demesne 129, 133 Duero (Spain) 179
denarii 263 basin 144–5, 160
Denmark 90 plateau (or northern plateau) 145,
Deobrigula (Tardajos, Burgos, Spain) 147, 160
161 River 143–5, 147, 160, 192
depopulation 260 Dundry Hill (Avon) 16
depopulation-repopulation historio- dyke 62
graphical model 143–9 Dynhagi (Iceland) 81, 83, 84
Dercinoassedenses (Spain) 202 Dýjakot (Iceland) 59
De rerum naturis 285 Dýrafjörður (Iceland) 93–4, 100
Desiderius, abbott 291
De situ orbis 285 early modern period 88, 92, 95–6, 102,
Destruction of sites 282 105
De universo 285 earthworks 46, 48, 55, 56, 57, 58, 60,
Devil’s Dyke (Cambridgeshire) 24 61, 62, 64
Page 368
Ordernr. 061297 P1
Index 353
Ebro valley (Spain) 145, 168, 170, 172, Fardagar (moving days) 75
174, 176, 178 farms
Ecgwine, bishop 125 names 58
Edgar, King of England 138 secondary farms (Icelandic) 59, 62, 63
Edict of Pîtres 40 sites 46, 52, 60, 61, 87, 91–3,
Edward the Elder, King of Wessex 242, 96–109, 111, 113
243 Farnham (Surrey) 36
Efri-Ás (Iceland) 105 Faroes 90–2, 100
Efri-Langahlíð (see also Langahlíð) 81 Fell (Iceland) 74
Efri-Rauðalækur, see Rauðalækur feorm 130
Eigil 284 Fernando I, King of León-Castile
Einarsstaðir (Iceland) 59 (Spain) 157–8, 173 (n.19)
elites 194 Fernán González, Count of Castile
Emerita (Spain) 198 (Spain) 178
enclosure 54, 56, 58, 62 Ferriéres, monastery (France) 285
endowment 109–10 feudalization 179, 184
English Combe (Avon) 15, 16 ‘Feudal Revolution’ 258
Ennery (Moselle, France), Merovin- field survey 45, 46, 52
gian cemetery 221, 222 Fines (bounds/limits) 216, 217, 218
environmental determinism 45
fishing 96
episcopal see 147, 160, 183
fishing station 96
Ermita del Monte (Spain) 173
Fiskilækjarhverfi (Iceland) 102, 106
Espinosa (Spain) 184
Fiskilækur (Iceland) 103
estates 26, 31, 151, 159
fission 182
‘great estates’ 257, 258, 272
Estepa, Carlos 159 Fjörður (Iceland) 97
ethnic identities 194 Flatagerði (Iceland) 59
Europe 143, 285, 286 Flatey (Iceland) 96
Evesham Abbey (Worcestershire) 125, Fleury, monastery (France) 283, 291
126, 127, 139 Fljót (Iceland) 71, 77
Exeter (Devon) 39, 241, 248, 251, 254 Fljótsheiði (Iceland) 62
Eyjafjörður (Iceland) 60, 61, 65, 66, 72, Fljótshverfi (Iceland) 51
73, 74, 77, 78, 93–4, 99–100, 103 Flói (Iceland) 99, 109
Eyjólfr Hallsson 73–4 Fnjóskadalur (Iceland) 71, 83, 85, 101
font 124, 125
Faccombe Netherton (Hampshire) Fora 189
241, 248 Foreigners 193, 198
factories 189 Fornastaðir (Iceland) 85
familia 263, 272 fort 167, 177, 178
family 87, 89, 105, 261, 264, 266, 267, fortifications 40, 168, 170, 173, 176,
272 179, 255, 266, 268, 270, 271
Page 369
Ordernr. 061297 P1
354 Index
fortress 159, 163, 180, 181, 182 Goði, pl. Goðar (chieftain), see Goðorðs-
Fosse Way (Avon) 17 maðr*
foundation legends 276, 278 (n.6) Goðorð, pl. goðorð (chieftaincy) 78
France 234, 248, 281, 284 Goðorðsmaðr, pl. Goðorðsmenn (chief-
Franks 284, 290 tain) 69, 74, 75, 78, 79, 82, 84, 86
freedom 263 Gog and Magog 288
Fritzlar, monastery (Germany) 275 Gorze (Moselle, France), abbey 217
frontier 14, 43, 44, 181 Cartulary of 217–8
Fulda (Germany) Goths 190
monastery 276, 284 grazing rights 146, 157–8, 165–6
river 275 Great Faringdon (Berkshire) 241, 248
Fundus 209 Greenland 90
Furufjörður (Iceland) 97 Gregorian reform 257
Fyfield Down (Wiltshire) 25 Gregory of Tours 280 (n.10)
Fylgðarmaðr (follower, supporter) 79 Grenjaðarstaðir (Iceland) 50, 51, 52,
fyrndarbýlingur 91 60, 63
Grim’s Ditch (Oxfordshire) 22
Galicia (Spain) 143 Grindavík (Iceland) 93, 95–7
García III, King of Navarre and Castile Grindill (Iceland) 77
173 (n.19), 181 Grímsey (Iceland) 96
Garðar (Iceland) 105 Grímsnes (Iceland) 99
Garðar Svavarsson 48 Grund (Iceland) 81
Garðshorn (Iceland) 74 Guðmundr dýri Þorvaldsson 65–86
Garður (Iceland) 103–4 Guðmundar saga dýra 65–86, 104
Gálmr (Grímsson) 84 Gufuskálar (Iceland) 103–4
Gásir (Iceland) 78 guilds 251
gathering 87–8, 107 Güdingen (Kreis Saarbrücken, Ger-
Geirrekur (Iceland) 51 many), Merovingian cemetery 221
Geitafell (Iceland) 51
Geldingaholt (Iceland) 73 Hadrian’s Wall 19
Gerald of Aurillac, saint 70 Hagi (Iceland) 105
Germany 284 Halldórsstaðir (Iceland) 59
gift 264 Hallfríðarstaðir (Iceland) 81
Gildas 31, 44 Hallr Gunnarsson 82
Glastonbury (Somerset) 37 Halwell (Devon) 241, 242
glebe 137 Ham (Wiltshire) 20, 21
Glerá (Iceland) 82 Hamburg-Bremen (Germany) 285, 286
Gloucestershire 37 hamlet 88–9, 91, 98–9, 103
Glæsibæjarhreppur (Iceland) 81 Hampshire 36, 41, 44
Glæsibær (Iceland) 82, 85 Hamtun (Hampshire) 238
Page 370
Ordernr. 061297 P1
Index 355
Hastings (Sussex) 120 Holtakot (Iceland) 52, 59

Hákon Þórðarson 75 homefield (see also infield) 49, 52, 54,
Hálfdanartunga (Iceland) 85 56, 58, 59, 62
Háls (Iceland) 71, 80 homefield boundary 52, 55, 56
head dyke 91 Honoratus, bishop 281
Heiðarbót (Iceland) 49, 50, 56, 59, 62 Honorius, Roman Emperor 190
Helgastaðir (Iceland) 73–4 Hontoria (Spain), territory 150–2,
Hellissandur (Iceland) 96 154, 164
Hellnar (Iceland) 96 Hontoria de la Cantera (Hontoria de
Henley-in-Arden (Warwickshire) Suso, Spain) 151–5, 162
122–4, 130 church of San Felices 154–5
Henry II, emperor 291 church of San Miguel 154
Heorot 40 Hontoria de Yuso (Spain) 151, 153
Herfeld (Germany) 275 church of Santa Eugenia 153–4
Hermitage 170, 171, 178 Hornbrekka (Iceland) 104
Héðinsfjörður (Iceland) 97 Hornstrandir (Iceland) 93–4, 97–9
hides 120, 134, 140 Horsecombe (Avon) 14, 15
Hierarchization 182 Down 17
hierarchy 171, 172, 180, 184 Vale 17
High Middle Ages 95 horse fighting 77, 88
high-status burial hospitium 191
Anglo-Saxon 33 household 89, 91, 93, 96, 99, 102, 105,
Merovingian 221–2, 224 107, 113
hillforts 170, 171, 172, 179, 195, size 88–9, 93, 96–8, 100–1
205–6, 242 Hólar (Iceland) 69, 72, 73
hilltop site 162–4, 167, 168, 170, 171, Hólsfjöll (Iceland) 99
172, 173, 174, 175, 176, 178, 179, Hraban Maur 285
180, 181, 182, 183, 184, 185 Hraun (Iceland) 105
hinterland, see alfoz, district, subur- Hreppstefna (meeting of a hreppur) 80
bium, territory Hreppur, pl. hreppar (commune)
historical archaeology 90 78–80, 81, 82, 86, 102
historiography 167 Hrísakot (Iceland) 58, 59, 62, 64
medieval Iberia 143–9 Hrothgar 40
medieval Iceland 67 Hugh of Hatton 130, 131, 132
Hjalli (Iceland) 108 hundreds, hundredal centres 13, 117,
Hjaltadalur (Iceland) 105 125, 135, 139, 142, 237, 238, 240, 251
Hjaltastaðir (Iceland) 108 Húnaþing (Iceland) 99
Hlaðir (Iceland) 81, 83 Húsavík (Iceland) 48, 50, 97
Hoba (farm) 218 Hvammar (Iceland) 59, 105
Holt (Iceland) 96, 99 Hvammsheiði (Iceland) 46, 57, 58
Page 371
Ordernr. 061297 P1
356 Index
Hvammur (Iceland) 105 Kaldaðarnes (Iceland) 101

hverfi 102 Kálfr Guttormsson 84
Höfðagerði (Iceland) 59 Kárastaðir (Iceland) 59
Höfðingi, pl. höfðingjar (leader) (see Kelduhverfi (Iceland) 51, 102
also Goðorðsmaðr) 69, 84 Kenilworth Priory (Warwickshire)
Hörgárdalur 65, 66, 70, 72, 75, 77, 78, 127, 128
79, 81–4, 85, 93–4, 99, 109 Kennet (Wiltshire) 32, 41
Höskuldur 48 Kent 36
Keynsham (Avon) 16
Ibeas (Spain) 157 king’s tûn 130
Iberia 144 Kington (Warwickshire) 130, 137
Iberian Peninsula 167, 184 kin relationships 112, 195
identity 105–6, 108, 112 Kirkjudagr (Iceland) 74, 77
Ijedo (Spain) 173 Klængr (Kleppjárnsson) 74
Ildeprand 290 (n.32) Klömbur (Iceland) 105
Illugi ( Jósepsson) 83–4 Koenigsmacker (Moselle, France),
industry 248 Merovingian cemetery 221
Ine, King of Wessex 36 Kolbeinn Tumason 71, 72, 73, 74, 78,
infield 91, 104 84, 86
Ingibjörg, daughter of Guðmundr dýri
Þorvaldsson 78 Lacock (Wiltshire) 17
inheritance 264, 265, 266, 272 La Horadada (Spain) 176
Inkpen Beacon (Berkshire) 20, 23, 24, landholding 151–2, 155
35, 36 landscape 89, 91–2, 100, 104, 105, 152,
Innnes (Iceland) 93, 95, 98–101, 107 155, 162–3, 168, 172, 174, 276, 280
Iona (Scotland), monastery 293 Landssveit (Iceland) 99
Iron Age 170, 171 land transactions 151–2, 155, 157, 165
Isidore of Seville 285 Langahlíð (including Skriða and Efri-
Italy 44, 276, 290 Langahlíð) (Iceland) 71, 75, 78, 81,
itineraries 37 82, 83, 84
Íradalur (Iceland) 54, 58 Langanes (Iceland) 97
Langholt (Iceland) 111–12
Járngerðarstaðahverfi (Iceland) 51 Langley (Warwickshire) 130–2
Járngerðarstaðir (Iceland) 98 Langmauer 213
Johannes 290 (n.32) Langport (Somerset) 241, 242, 252
John of Bearley 128, 129 language and space 293
Jonas of Bobbio 279 Lantarón (Spain) 178, 180
Jón Loptsson 71 Lara de los Infantes (Spain) 161, 165,
justice 158–9, 166 196, 199
alfoz 165
Page 372
Ordernr. 061297 P1
Index 357
La Rioja (Spain) 168 longhouse 58

Las Ermitas (Spain) 172 Long Newton (Wiltshire) 37
late antiquity 167, 170 Lora (Spain) 173
Laufás (Iceland) 75, 85 Lord 178
Laugaland (Iceland) 70, 72, 77, 81, 82, lordship, lords 118, 129, 178 (n.33),
83, 84 180, 182, 183, 184, 257, 266, 267
Laugardælir (Iceland) 109, 111–12 ecclesiastical 152, 158
Laugarnes (Iceland) 107 large seigneurial dominions 147
Lavoye (Meuse, France), Merovingian Losa (Spain) 172, 182
cemetery 220, 221, 222, 224 Louis I ‘the Pious’, King of the Franks
law, laws 266, 268 215, 285
Grágás (Iceland) 79 Louis II, Emperor of the Franks 280,
Ripuarian 229 291
Salic, see Pactus Legis Salicae Lækur (Iceland) 111
trading 243, 244 lögbýli 91, 93
Laxamýri (Iceland) 48, 49, 50, 52, 60,
62, 64 Macrobius 280 (n.10)
Laxá (Iceland) 46, 49 Maes Knoll (Avon) 14, 15, 16, 21, 24,
Leið (Icelandic autumn assembly) 80 36
Leira (Iceland) 104 Malmesbury (Wiltshire) 36
Leo of Ostia 291 abbey 37
León (Spain) 188 house 118, 126, 127, 128, 129, 135,
kingdom 147 138, 142
kings of 144–5 manor 117, 118, 119, 125, 127, 129,
Lerma (Spain) 161 131, 132, 133, 134, 137, 139, 142
Life of Romaric 227 Mansiones 189
Lindisfarne (Northumberland), mon- Mansus (farm) 218
astery 293 maps 116, 122
Lindsey family 132 Marca (bounds/limits) 216
Lisbon (Portugal) 202 Mardancho (Spain) 173
Litla-Saltvík (Iceland) 58, 59 Maríugerði (Iceland) 52, 59
Litla-Þverá (Iceland) 59 marketplaces 189
Litlu-Reykir (Iceland) 59 Marlborough (Wiltshire) 17, 25, 41,
Litlu-Skörð (Iceland) 54, 59 241, 242
Ljósvetninga saga 49, 50 Mass 87
local citizenship 202 Mataviejas (Spain), River 164
local communities 189, 193, 256, 259 Mave (Spain) 170, 178
Lombards 282 medieval maps 288
London (England) 241, 244, 252 Medina de Pomar (Spain) 170, 176, 182
Long Ashton (Somerset/Avon) 16 Mediterranean sea 285
Page 373
Ordernr. 061297 P1
358 Index
meeting places 304 Modúbar de la Cuesta (Spain) 151

Melar (Iceland) 105 Modúbar de la Emparedada (Spain) 151
Melasveit (Iceland) 105 monasteries 170, 175, 179, 181, 184
memory 282, 289–93 monastic community 119, 129, 134,
Mercia, kingdom of 23, 32, 35, 36, 37, 139, 278, 279, 280, 281, 282, 286,
40, 44, 126, 139 291, 293
Merovingian Cemeteries monastic rules 289
distribution of 214–15, 224, 228 Mondsee (Austria), monastery 276
organization of 222, 224 moneyers 244, 250, 251
Meseta (Iberian plateau) 178, 179 Montecassino (Italy), monastery 282,
sub-meseta 260 283, 291
metalworking 248, 249 Monte Cildá (Spain) 170
Metz 220, 222, 224, 226 Montfort family 135
region of 209, 212, 218, 219, 220, 226 Moradillo del Castillo (Spain) 182
Métrich (Moselle, France), Merovin- Morgans Hill (Wiltshire) 17, 18, 41
gian cemetery 221 Morton Bagot (Warwickshire) 127–8,
Miðhvammur (Iceland) 59, 104–5 130
Miðskálagarður (Iceland) 103–4 mother-church 115, 117, 118, 119, 122,
Mijangos (Spain) 176, 178 (n.33) 124, 125, 127, 128, 129, 130, 132,
Mildenhall (Wiltshire) 17 133, 134, 138, 139, 142
military obligations 39, 158–9, 166 mountains
milites 261, 266, 268 Iberian 144
mining sites 189 Pyrenees 144
‘ministerial serfs’ 263, 271 Spanish Central 143
minster hypothesis debate 138, 140 movement of serfs 264, 272
parishes of 118, 119, 120, 125, 127, Mozarabic 175
132, 137, 138, 139, 140, 142, 238 Mudá (Spain) 183
minsters 117, 119, 120, 134, 135, 138, Muñó (Spain) 161
140, 142, 237, 239, 242, 254 Muslim 146, 159
mint, mints 240, 244, 253 Músarhóll (Iceland) 111
Miracula St Adalardi 278 Myrkárdalur (Iceland) 81, 83
Miranda (Spain) 177, 179 Mýrar (Iceland) 99
mobility, spatial 3–4, 7, 299 Mýrarsel (Iceland) 59, 60
Modúbar (Spain) Mýrarvatn (Iceland) 49, 60
territory 150–2, 155, 164–5 Mýri (Iceland) 49, 58, 59
valley 151, 164 Mývatn (Iceland) 56, 61
Modúbar de San Cebrián (=Modúbar Mývatnssveit (Iceland) 62, 90
de San Cibrián, Spain) 151, 155 Mælifell (Iceland) 105
church of San Cebrián 151 Möðruvellir (Iceland) 70, 71, 77, 80,
Modúbar de Zahalanes (Spain) 151–2 81, 82, 83, 84
Page 374
Ordernr. 061297 P1
Index 359
Nanteuil (France) 255 Olleros de Pisuerga (Spain) 170

Naveil (France), church at 261–2, 270 Omecillo (Spain) 171
Náttfaravíkur (Iceland) 97 open fields 135
nearest neighbour 93, 97–8, 101 oppidum 160, 161, 194
Nebreda (Spain) 164 Orbaneja (Spain) 149, 155, 158, 165
Neðri-Ás (Iceland) 105 area 157
neighbourhood 98, 103–4, 106, 111–12 monastery of San Vicente 157–8
Nennig (Triererland, Germany), villa valley 157–8
214 Orbaneja de Picos (Spain) 156–8
Nes (Iceland) 107 ordeal of hot iron 265
New Buildings (Wiltshire) 25, 41 Osma (see also Uxama, Spain) 161
Newbury (Berkshire) 241, 248 Osric, King of the Hwicce 36
Newton St Loe (Somerset/Avon) 16 Otto, bishop 281
Nikulás Skratta-Bjarnarson 77–8 outfield 46, 91
nobility 234, 261 outsiders 2, 303
Nonantola (Italy), monastery 282 (n.13) Overton Down (Wiltshire) 25, 42
Norman Conquest, of England 246, 249 Oviedo (Spain), kings of 144
Norse 91 ownership (Spain) 172, 173, 175
North Atlantic 90 Oxford (Oxfordshire) 241, 246, 247,
Northern Isles 90 248, 251, 252
Norton Lindsey (Warwickshire) 130–3, Ófeigur Járngerðarson 49
140 Ólafsfjörður (Iceland) 71, 104
Norway 87–8, 90, 92 Ósbrekka (Iceland) 104
Nova Augusta (Spain) 196 Ósgerði (Iceland) 111
Núpar (Iceland) 107, 109
Pactus Legis Salicae 227, 229
oblivion 282, 289–92 pagan graves 50
Oca (Spain) 176 pagi 189, 205, 290 (n.32)
ocean 285 Merovingian 226
Odd Down (Avon) 17 Pagus Salinensis 216
Odo, Prior of Marmoutier 262, 265, Paldo 283
267, 269 Palencia (Spain) 168, 183
Offa, King of Mercia 34, 37, 40 Panero (Spain) 173
Offa’s Dyke 16, 18, 19, 24, 31, 34, 41, 43 Paredes Rubias (Spain) 181
Okehampton (Devon) 241, 248 parish churches 50, 63, 89, 100, 104,
Oldaport (Devon) 242 107–8, 111, 115, 118, 119, 131, 132,
Oldberrow (Warwickshire) 125–7, 128, 257
130, 138–9 parish system 26, 89, 91, 102, 105–9,
Old Sarum (Wiltshire) 22, 241, 242 111, 112, 115, 117, 118, 119, 120,
Old Shepherd’s Shore (Wiltshire) 23 122, 124, 138, 142, 173
Page 375
Ordernr. 061297 P1
360 Index
Anglo-Saxon (England) 149 placitum 266

Castilian (Spain) 147, 149, 159, 166 plebs 6–7, 8, 9, 298
Icelandic 51, 76–77, 80, 86 polarization of space 287 (n.22)
priests 257 political identities 193, 197–200
registers 124 politicization 4–5
Pas-de-Calais (France) 281 polity 178, 179, 184
pastoral care 115, 118, 119, 125, 139 polyptychs 272
pasture (see also grazing rights) 146, population 68, 90–1, 93, 98, 105, 113
152, 157–8 density 6
patronage 184 Portbury Camp (Avon) 16
Patronus 197, 204 Portishead (Avon) 16
peasant 172, 176, 180, 182, 234 ports 244
Peña de Nuestra Señora (Spain) 206 pottery 170, 171, 238, 249, 253, 254
Peñas Negras (Spain) 183 chaff-tempered 42
pension 137 imported 244, 248
Peralejo de los Escuderos (Spain) 200 scatters 8
Peregrini 193, 198 workshops 189
Pernía (Spain) 183 power and space 89, 276, 292–4
Pescara (Spain), River 280 preaching cross 125
Peter de Mora 128 pre-Roman periods 10, 89, 168, 171,
Petronax, abbott 282 174, 183
Phoebus 281 Presthvammur (Iceland) 49, 104–5
Picos (Spain), River 155–6 Preston Bagot (Warwickshire) 130,
Piedralada (Spain) 182 133–7
Piedras Negras (Spain) 183 priests 118, 119, 124, 125, 132, 134,
Pilton (Devon) 241, 242 135
Pisuerga (Spain) 168, 170, 183 primary farms (Icelandic) 59
River 144 priories 127, 128, 269–70, 273
place-names 50, 51, 60, 89–90, 92, Priston (Avon) 16
102–6, 109, 112, 122, 133, 146, ‘Productive sites’ 238
149, 151–2, 154–5, 164–5, 170, Publow Hill (Somerset) 15
174, 180, 181, 183 puritanism 88
-gerði (Icelandic place-name element) Puxieux (Meurthe-et-Moselle, France),
59 Merovingian cemetery 221
-hverfi (Icelandic place-name element)
50, 51, 52, 63 Quentovic (France) 40
-kot (Icelandic place-name element)
59 Rainaldus, abbott 290
-staðir (Icelandic place-name element) Ramsbury (Wiltshire) 37, 246
59 Rangárvellir (Iceland) 99
Page 376
Ordernr. 061297 P1
Index 361
Rangárþing (Iceland) 99 Rome 282, 283, 290

Rauðalækur (Iceland) 81, 83, 84 Roundway Down (Wiltshire) 33
Ratramnus 286 routes 172, 173, 176, 178, 183
Reading (Berkshire) 120, 251, 252 Icknield Way 20
Realengo (Spain) 184 Ridgeway 20, 31, 32
rector 131, 132 royal
Red Shore (Wiltshire) 20, 25 administrative centres 237, 238
regio 120 estate centres 37, 242, 248, 254
relics 280, 281, 282 (n.12), 283, 284, forests 287
291 residence 130
Resoba (Spain) 183 Ruanales (Spain) 172, 182
Revilla (Spain), territory 150, 164 Rubena (Spain) 157
Revilla de la Fuente (Spain) 151 Rudrón (Spain) 170, 181
Revilla del Campo (Spain) 151–2, 155, Rule of Benedict 278 (n.7)
162 Rybury (Wiltshire) 18
Reykdæla saga 49, 50, 51, 105
Reykir (Iceland) 48, 49, 50, 51, 52, 58, S. Andrew on Monte Soratte (Italy),
59, 60, 62, 63, 107–9 monastery 290
Reykjadalur (Iceland) 59, 73 S. Clemente a Casauria (Italy), monas-
Reykjahreppur (Iceland) 51 tery 280, 281, 290, 291
Reykjahverfi (Iceland) 45, 46, 47, 48, 50, S. Clemente in Rome, basilica 291
51, 52, 56, 57, 59, 60, 61, 62, 63, 102 S. Salvatore alla Maiella (Italy) 290
Reykjasel (Iceland) 60 S. Vaast (France), monastery 281, 282
Reykjavellir (Iceland) 111 S. Vincenzo al Volturno (Italy), mon-
Reykjavík (Iceland) 107 astery 283
Rhineland 248, 298 Sacralization of space 287, 288–9
Rif (Iceland) 96 Saelices (Spain), River 153–4
Rimbert 279 (n.8), 286 (n.20) sagas (Icelandic) 45, 46, 48, 49, 50, 51,
Rimo 290 (n.32) 56, 77, 80
ringwork 127, 135 Saltford (Avon) 16
Roa (Spain) 161 saltpans 189
Rodings, the (Essex) 120 Saltvík (Iceland) 49, 50, 58, 59, 62
Roman 168, 170, 171, 172, 175, 176, Saltvíkurgerði (Iceland) 59
177, 185 San Felices de Bilibio (Spain) 178
citizenship 194, 196 San Martín de Elines (Spain) 175
decadence of 188 San Martín de Lantarón (Spain) 178
inscriptions 188, 193, 197–9, 201 San Miguel de Infravillas (Spain) 184
municipia 194 (n.59)
Romance of Alexander 288 San Millán de la Cogolla (Spain) 204
Romanization 188
Page 377
Ordernr. 061297 P1
362 Index
San Pedro de Cardeña (Spain), abbey, central 159–60, 162–3, 166

monastery 155–8 chronology 91, 95, 102
San Salvador de Cantamuda (Spain) dependent 147, 158
184 (n.59) dispersed 146–7
San Salvador de Oña (Spain) 170 farm-type 146
Sandar (Iceland) 100 hierarchy 54, 89, 147–8, 159
Sandur (Iceland) 79 local 147
Sandvík (Iceland) 97 medieval Castile-León (Spain) 145–8
Sanson 290 (n.32) names 59
Santa Gadea del Cid (Spain) 170, 180 nucleated 164–5, 182, 257, 272–3
Santa María de Hito (Spain) 175 patterns 41, 45, 46, 63, 88, 90–2,
Santa María de Lebanza (Spain) 184 95–101, 108–9, 111, 147–9, 155,
(n.59) 189, 214–15
Santa María de Mave (Spain) 170 planned 99–100, 109, 111–13
Santa María de Mijangos (Spain) 176 rural 147–8, 156
Santas Centola y Elena (Spain) 170 Settlements, The (Icelandic) Book of
Savernake Forest (Wiltshire) 18, 20, 45, 46, 48
23, 25 Severn, River 15, 16, 21
Saxons 32, 275 Severus 31
Sánchez-Albornoz, Claudio 145 Shalbourne (Wiltshire) 20
Scandinavia 90, 279 Shetland 91
Scarponne (Meurthe-et-Moselle, shielings 46, 51, 60–61
France) 226 shires, shire centres 237, 244
Scholastica 283 Short Oak Copse (Wiltshire) 20
seasonality 99 Siero (Spain) 170, 181
Sedano (Spain) 181 Sigillata 171
Segobriga (Spain) 192 Silbury Hill (Wiltshire) 42
Selfoss (Iceland) 111 Silchester (Hampshire) 23, 33
Selvogur (Iceland) 96 Sisenandus 290 (n.32)
Selwood Forest (Somerset/Wiltshire) Skagafjörður (Iceland) 73, 74, 85, 105
18 Skarðagerði (Iceland) 52, 54, 58, 59, 62
Sermo de relatione corporis Beati Vedasti Skarðasel (Iceland) 60, 61
282 (n.12), 284 Skeggjabrekka (Iceland) 104
servile status 262–3, 264, 265 Skeið (Iceland) 99
servitude 255, 256, 263, 268, 271 Skjálfandi (Iceland) 46, 48
Servus 197 Skógar (Iceland) 49, 51, 59
settlement 45, 46, 48, 49, 50, 51, 58, Skriða (in Aðaldalur, Iceland) (see also
59, 60, 61, 62, 63, 88–90, 96–7, Langahlíð) 51, 81 (n.39)
112, 255, 256 Skriðuhverfi (Iceland) 51
abandonment 58, 60, 63 Skriðuþingsókn (Iceland) 81–2
Page 378
Ordernr. 061297 P1
Index 363
Skörð (Iceland) 48, 49, 50, 51, 52, 54, Sturm 275, 284
58, 59, 60, 62, 64 suburbium (see also alfoz, district, terri-
Smjördalir (Iceland) 111 tory) 157, 161
Snitterfield (Warwickshire) 129 Sueves 190
social network analysis 67 Sumarliði (Ásmundarson) 74
social strucure, Iceland 69, 89, 90 supralocality 13, 301, 302
solidi 262 Surrey 36, 42
Somerford Keynes (Gloucestershire) 37 Svarfaðardalur (Iceland) 71, 73, 74, 80
Somerset 31, 32, 35, 41, 251, 254 Svarfhóll (Iceland) 109, 111
Somerton (Somerset) 238, 241, 242 sveit (Iceland) 102
Soria (Spain) 200 Swallowcliffe Down (Wiltshire) 33
Sotoscueva (Spain) 184 Sweden 90, 279
sources, primary 1 Sæmundr Jónsson 85
South Cadbury (Somerset) 241, 242 Sölfholt (Iceland) 111
South Stoke (Avon) 14
sókn 100, 106 Tabladillo (Spain), alfoz 164–6
sóknarkirkja 100 Tabladillo (de Lara, Spain) 165
Spain, medieval 144 Tabula patronatus 197, 204
spiritual privileges 261, 263, 266 Tabular Hills (Yorkshire) 21
Spye Park (Wiltshire) 24 Talamanquilla (Spain) 164
St Florent’s Abbey, Saumur (Anjou, Tan Hill (Wiltshire) 18
France) 130, 132 Taso 283
St Martin 259 Tato 283
St-Ulrich (Moselle, France), villa 212 taxation
Staðarhverfi (Iceland) 98 Icelandic 91–2, 106, 108, 111
Staðr (church farm) 73 Merovingian 227–9
Staður (Iceland) 98 tax registers: Jarðabók Árna Magnús-
Stantonbury (Avon) 15, 16, 18 sonar og Páls Vídalíns (Iceland) 65,
states 88, 233, 235, 236, 244, 250 68, 78, 80
status 106 Tedeja (Spain) 176, 178, 180, 182
Steinsstaðir (Iceland) 81, 83 tephra 50, 58
stekkur (a weening fold) 61 Termes (Spain) 200
Stekkur (Iceland) 58, 59 territorial expansion
Steyning (England) 239, 240, 241 Asturian (Spain) 144
Stokkseyri (Iceland) 96 Carolingian144
Stoppingas 120, 142 urban 157
Stratford-upon-Avon (Warwickshire) village 149, 151–2, 155, 157–8
133 territoriality 147, 152, 157, 159, 166
Streatley (Berkshire) 36 animal 2
Sturlungar 85 human 2–3, 4, 5, 302
Page 379
Ordernr. 061297 P1
364 Index
territory (see also alfoz, district, subur- Ullenhall (Warwickshire) 122, 124–5,
bium) 147–8, 150, 155, 159, 160, 130, 134
163, 165, 168, 171, 172, 174, 175, uncultivated land, management of 146,
178, 179, 180, 181, 183, 184 158, 165
articulation 175, 181 uplands 183, 184
fission 149–50, 154–5 Uppsalir (north Iceland) 85
organization 172, 173, 174, 184 Uppsalir (in Flói, Iceland) 109, 111
supralocal 147–8, 149, 151–2, 154–5, Ura (Spain), alfoz 164
157, 158–60, 165–6, 171, 179, 181, urban (places) 87–9
183 buildings 250
Tesla (Spain) 176 development 192
Tetbury (Gloucestershire) 37 fields 251
Término (Spain) 180, 182 hinterlands 249, 250
Thames, River 42, 44 identity 233, 237, 250
thegns 246, 248 manors 246
Theudebert II, King of Austrasia 228 urbanization 42, 179, 233, 234, 240,
Theuderic II, King of Burgundy 227–8 chronology of 235, 236, 254
tithes, tithings 69, 106–8, 131, 133 in the twelfth century 249
Tjarnargarðshorn (Iceland) 77 Uxama (Burgo de Osma, Soria, Spain)
Tjörn (Iceland) 74, 77 160, 192, 193
Tjörnes (Iceland) 60, 61, 62 Uxastaðir (Iceland) 54, 59
Tjörvi (from Rauðalækur) 84 Úthérað (Iceland) 93–4, 99, 108
Tobalina (Spain) 178 Útskálar (Iceland) 103–4
Toledo (Spain) 178
topography 149, 152, 154, 157–8, 163–5 Vadinienses (Spain) 188
Tours (France) 255, 259, 263 Vaðalsþorp (Iceland) 102
tower 176 Vaðlaþing (Iceland) 71
trade, trading 234, 245, 248, 253 Vagn 133
transhumance 172, 184 Vandals 190
Translatio Sancti Benedicti 283 (n.14) Valdebodres (Spain) 184
Translatio Sancti Clementis 291 Valdegovía (Spain) 171, 178, 179
transport 97, 99, 101–2 Valdelaguna (Spain) 166
Tremaya (Spain) 183 Valdelomar (Spain) 173
tribute 177 Valderredible (Spain) 172, 173, 174,
Trier (Germany) 212, 213 181
Trinoda Necessitas 40 Vale of Pewsey (Wiltshire) 18
Tritium (monasterio de Rodilla, Burgos, Valeria (Spain) 192
Spain) 161 Vallejo de Santillán (Spain) 176, 180
Turholt, monastery (Belgium) 286 várþing (Modern Icelandic vorþing,
Tún (Iceland) 111 spring assembly) 80
Page 380
Ordernr. 061297 P1
Index 365
Vellir (Iceland) 73 Vita Sturmi 275–6, 284

Venta Belgarum (Hampshire) 21 Víga-Glúms saga 80 (n.36)
Verlucio (Sandy Lane, Wiltshire) 26, 42 Víkingsstaðir (Iceland) 59
Vestfirðir (Iceland) 97 Votmúli (Iceland) 111
Vestmannaeyjar (Iceland) 96 Vöðu-Brandur 49
vicani 227
vicar 124 Waldhambach (Bas-Rhin, Germany),
vici 189, 205, 220, 255 villa 216
Vicus Spacorum (Spain) 198 Wallingford (Oxfordshire) 39, 241, 245,
Vigo (Spain) 198 248, 249, 251, 252
Viking age, vikings 32, 88, 90, 95, 97, Wansdyke
102, 107, 260, 281, 285 context of 26–38
villa, Merovingian 214–31 dating of 21–5
communities of 220–6 warfare 32
definition of 215–19 Wareham (Dorset) 241, 249, 252
royal grants of 227, 229, 230 Warwickshire 125, 134, 138
and taxation 227–9 water, access to 147, 157, 165
villa, Roman 209–14 watersheds 152, 163–4
abandonment of 210, 212, 220 Wends 279
estate 172, 175, 179 Wells (England) 246
Villa de Domna Eilo (Spain) 156, 158 Wernham Farm (Wiltshire) 25
Villa de Plano (Spain) 156, 158 Wessex, Kingdom of 23, 32, 33, 34, 35,
Villa de Valle (Spain) 156, 158 36, 37, 39, 44, 235, 236, 237, 245,
villae 189, 255 246, 253, 254
Villafría (Spain) 157 Westphalia (Germany) 275
villages 26, 87–8, 91, 96–7, 100, 101–2, West Woods (Wiltshire) 18, 19
147, 149, 151–5, 157–8, 164–6, 170, Whitley (Warwickshire) 134
173, 174, 181, 182, 183, 256, 272, Wigbert, abbott 275
273, 298–9, 303 William of Bearley 128, 129
Villalba de Losa (Spain) 182, 183 William de Goudulph 17
Villanueva la Nía (Spain) 173 Wiltshire 31, 35, 36, 37, 41, 44, 251
Villarcayo (Spain) 176 Winchester (Hampshire) 235, 241,
Vindeleia (Spain) 198 243, 245, 246, 248, 251, 252, 254
Visigothic 172, 176, 178, 179 Windmill Hill (Avon) 16
collapse 143 Wissembourg (Bas-Rhin, France), abbey
institutional inheritance 145 216, 276
Vita Anskarii 279 (n.8), 286 cartulary of 216–17, 218, 230
Vita Columbani 279, 280 (n.9) woddesgeat (Wiltshire) 32
Vita Rimberti 286 (n.20) Woden 21, 32, 33, 34
Vita Sancti Venantii 280 (n.10)
Page 381
Ordernr. 061297 P1
366 Index
wodnesbeorg (Adam’s Grave, Wiltshire) Æthelbald, King of Mercia 120, 140

31, 32, 33 Æthelmund, minister 40
wodnesdene (Wiltshire) 32 Æthelwulf, King of Wessex 37, 40
Wolverton (Warwickshire) 130–2
women 87 Ögmundr Þorvarðsson 71, 73, 83, 85
wood, access to 147, 165 Ölfus 93–4, 99, 100
Wootton Wawen (Warwickshire) Öndólfsstaðir (Iceland) 59
119–42 Önundarbrenna (Iceland) 71, 86
Worcester (Worcestershire) 120, 122, Önundr Þorkelsson 70, 71, 73, 74, 78,
125, 126, 129, 131, 135 82, 83, 84
Worcestershire 125, 138, 139 Öxnadalur (Iceland), see Hörgárdalur
written sources 89 Öxnahóll (Iceland) 74, 77, 81, 82, 83
Yecla, La (Spain) 206

Yeovil (Somerset) 241, 248
Ystihvammur (Iceland) 59, 104
Ytri-Skörð (Iceland) 49
Zahalanes (Spain), territory 152
Þegjandadalur (Iceland) 60
þing 99
Þingeyrar (Iceland) 73
Þingmaðr, pl. þingmenn 70, 77, 78, 79
Þistilfjörður (Iceland) 99
Þorfinnr Önundarson 78
Þorgrímr alikarl 71, 72, 73, 83, 84
Þorkell 49
Þorleifskot (Iceland) 111
þorp 102
Þorvarðr Þorgeirsson 70–1, 73, 74, 79,
82, 83, 85
Þórðr Þórarinsson 79
Þórir Bárðarson 83
Þórkötlustaðir (Iceland) 98
Þverá (Iceland) 49, 50, 51, 59, 62
River 48
Þykkvibær (Iceland) 96
Page 382
Ordernr. 061297 P1
S TUDIES IN THE E ARLY M IDDLE A GES
All volumes in this series are evaluated by an Editorial Board, strictly on academic grounds,
based on reports prepared by referees who have been commissioned by virtue of their special-
ism in the appropriate field. The Board ensures that the screening is done independently and
without conflicts of interest. The definitive texts supplied by authors are also subject to review
by the Board before being approved for publication. Further, the volumes are copyedited to con-
form to the publisher’s stylebook and to the best international academic standards in the field.
Titles in Series
Cultures in Contact: Scandinavian Settlement in England in the Ninth and Tenth Centuries,
ed. by D. Hadley and J. Richards (2000)
On Barbarian Identity: Critical Approaches to Ethnicity in the Early Middle Ages, ed. by
Andrew Gillett (2002)
Matthew Townend, Language and History in Viking Age England: Linguistic Relations
between Speakers of Old Norse and Old English (2002)
Contact, Continuity, and Collapse: The Norse Colonization of the North Atlantic, ed. by J. H.
Barrett (2003)
Court Culture in the Early Middle Ages: The Proceedings of the First Alcuin Conference, ed. by
C. Cubitt (2003)
Political Assemblies in the Earlier Middle Ages, ed. by P. S. Barnwell and M. Mostert (2003)
Wulfstan, Archbishop of York: The Proceedings of the Second Alcuin Conference, ed. by Matthew
Townend (2004)
Borders, Barriers, and Ethnogenesis: Frontiers in Late Antiquity and the Middle Ages, ed. by
Florin Curta (2006)
John D. Niles, Old English Enigmatic Poems and the Play of the Texts (2006)
Page 383
Ordernr. 061297 P1
Teaching and Learning in Northern Europe, 1000–1200, ed. by Sally N. Vaughn and Jay
Rubenstein (2006)
Narrative and History in the Early Medieval West, ed. by Elizabeth M. Tyler and Ross
Balzaretti (2006)
In Preparation
John D. Niles, Old English Poetry and the Social Life of Texts
The Crisis of the Oikoumene: The Three Chapters and the Failed Quest for Unity in the Sixth-
Century Mediterranean, ed. by Celia Chazelle and Catherine Cubitt
Page 384
Ordernr. 061297 SEM15 QPag (P64<>P65= folio 48<>49) = Na Katern 2
Plate 1. Hrísakot ruins in the foreground, marked by a double earthwork. Above

middle is the modern farm of Skörð. The straight earthwork in the foreground
marks the boundary between Laxamýri and Skörð. Photo: Árni Einarsson.
Page 1 of 2
Ordernr. 061297 SEM15 QPag (P64<>P65= folio 48<>49) = Na Katern 2
Page 2 of 2

(Studies in The Early Middle Ages, 15) Wendy Davies, Guy Halsall, Andrew Reynolds - People and Space in The Middle Ages, 300-1300-Brepols (2006)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Studies in The Early Middle Ages, 15) Wendy Davies, Guy Halsall, Andrew Reynolds - People and Space in The Middle Ages, 300-1300-Brepols (2006)

Uploaded by

Copyright:

Available Formats

Ordernr.

P EOPLE AND SPACE

STUDIES IN THE EARLY MIDDLE AGES

Editorial Board under the auspices of the

Elizabeth M. Tyler (University of York)

P EOPLE AND SPACE

Wendy Davies, Guy Halsall and Andrew Reynolds

British Library Cataloguing in Publication Data

© 2006, Brepols Publishers n.v., Turnhout, Belgium

All rights reserved. No part of this publication may be reproduced,

Printed in the E.U. on acid-free paper

List of Figures vii

List of Tables xii

Note on Glossary Entries xiii

Introduction: Community Definition and Community Formation 1

Social Identities on the Macro Scale: A Maximum View of Wansdyke 13

Settlement Organization and Farm Abandonment: 45

Geography, Communities and Socio-Political Organization 65

Communities of Dispersed Settlements: Social Organization 87

Boundaries of Knowledge: Mapping the Land Units 115

Mapping Scale Change: Hierarchization and Fission in Castilian 143

Central Places and the Territorial Organization of Communities: 167

The Ending of the Roman City: 187

Villas, Territories and Communities in Merovingian Northern Gaul 209

Community, Identity and the Later Anglo-Saxon Town: 233

Marmoutier: Familia versus Family. The Relations 255

Narrating Places: Memory and Space in Medieval Monasteries 275

Populations, Territory and Community Membership: 295

Cover Illustration: Arnarklettur (‘Eagle’s rock’), a natural marker on a boundary

Figure 10.1, p. 211. The Region of Metz.

Photo 1.1, p. 9. The contributors at Lara, northern Spain.

Table 3.1, p. 59. Settlement names in Reykjahverfi.

N OTE ON G LOSSARY E NTRIES

Where a word is highlighted in Bold in the following text, followed by an asterisk,

AM The Árni Magnússon Institute, Reykjavík.

INTRODUCTION : C OMMUNITY D EFINITION

pre-existing communities? Or did the needs of government in effect call local

monasteries like Redon or Landévennec or of the noble retinues of counts and

Photo 1.1. The contributors at Lara, northern Spain.

S OCIAL IDENTITIES ON THE M ACRO S CALE:

Andrew Reynolds and Alex Langlands

14 Andrew Reynolds and Alex Langlands

individual communities to spatial dynamics and territorial definition, collective

Wansdyke’s westerly termination is generally considered to be the hill-fort Maes

SOCIAL IDENTITIES ON THE MACRO SCALE 15

Figure 2.1. The course of Wansdyke in relation to shire boundaries

16 Andrew Reynolds and Alex Langlands

SOCIAL IDENTITIES ON THE MACRO SCALE 17

de Goudulph to Adam de Cloptone of a cottage with adjoining land in Aystone

The Roman Road from Bath (Aquae Sulis) to Mildenhall (Cunetio)

Opinion is divided on the course of Wansdyke from Horescombe Down to

18 Andrew Reynolds and Alex Langlands

SOCIAL IDENTITIES ON THE MACRO SCALE 19

20 Andrew Reynolds and Alex Langlands

The Bedwyn Dykes

SOCIAL IDENTITIES ON THE MACRO SCALE 21

Dating Wansdyke: A Review of the Evidence

22 Andrew Reynolds and Alex Langlands

SOCIAL IDENTITIES ON THE MACRO SCALE 23

24 Andrew Reynolds and Alex Langlands

SOCIAL IDENTITIES ON THE MACRO SCALE 25

26 Andrew Reynolds and Alex Langlands

Territory, Warfare and Politics: The Context of Wansdyke